🔧 chore: Comment out unused font family configurations and typography plugin in Tailwind CSS config

# Conflicts:
#	client/package.json
#	client/vite.config.ts
#	package-lock.json

.devcontainer/Dockerfile

@@ -0,0 +1,5 @@
+FROM node:18-bullseye
+
+RUN useradd -m -s /bin/bash vscode
+RUN mkdir -p /workspaces && chown -R vscode:vscode /workspaces
+WORKDIR /workspaces

.devcontainer/devcontainer.json

@@ -1,58 +1,18 @@
-// {
-// 	"name": "LibreChat_dev",
-// 	// Update the 'dockerComposeFile' list if you have more compose files or use different names.
-// 	"dockerComposeFile": "docker-compose.yml",
-// 	// The 'service' property is the name of the service for the container that VS Code should
-// 	// use. Update this value and .devcontainer/docker-compose.yml to the real service name.
-// 	"service": "librechat",
-// 	// The 'workspaceFolder' property is the path VS Code should open by default when
-// 	// connected. Corresponds to a volume mount in .devcontainer/docker-compose.yml
-// 	"workspaceFolder": "/workspace"
-//     //,
-// 	// // Set *default* container specific settings.json values on container create.
-// 	// "settings": {},
-// 	// // Add the IDs of extensions you want installed when the container is created.
-// 	// "extensions": [],
-// 	// Uncomment the next line if you want to keep your containers running after VS Code shuts down.
-// 	// "shutdownAction": "none",
-// 	// Uncomment the next line to use 'postCreateCommand' to run commands after the container is created.
-// 	// "postCreateCommand": "uname -a",
-// 	// Comment out to connect as root instead. To add a non-root user, see: https://aka.ms/vscode-remote/containers/non-root.
-// 	// "remoteUser": "vscode"
-// }
 {
-    // "name": "LibreChat_dev",
-    "dockerComposeFile": "docker-compose.yml",
-    "service": "app",
-    // "image": "node:19-alpine",
-    // "workspaceFolder": "/workspaces",
-    "workspaceFolder": "/workspace",
-    // Set *default* container specific settings.json values on container create.
-    // "overrideCommand": true,
-    "customizations": {
-        "vscode": {
-            "extensions": [],
-            "settings": {
-                "terminal.integrated.profiles.linux": {
-                    "bash": null
-                }
-            }
+  "dockerComposeFile": "docker-compose.yml",
+  "service": "app",
+  "workspaceFolder": "/workspaces",
+  "customizations": {
+    "vscode": {
+      "extensions": [],
+      "settings": {
+        "terminal.integrated.profiles.linux": {
+          "bash": null
         }
-    },
-    "postCreateCommand": "",
-    // "workspaceMount": "src=${localWorkspaceFolder},dst=/code,type=bind,consistency=cached"
-
-    // "runArgs": [
-    //     "--cap-add=SYS_PTRACE", "--security-opt", "seccomp=unconfined",
-    //     "-v", "/tmp/.X11-unix:/tmp/.X11-unix",
-    //     "-v", "${env:XAUTHORITY}:/root/.Xauthority:rw",
-    //     "-v", "/home/${env:USER}/.cdh:/root/.cdh",
-    //     "-e", "DISPLAY=${env:DISPLAY}",
-    //     "--name=tgw_assistant_backend_dev",
-    //     "--network=host"
-    // ],
-    // "settings": {
-    //     "terminal.integrated.shell.linux": "/bin/bash"
-    // },
-    "features": {"ghcr.io/devcontainers/features/git:1": {}}
+      }
+    }
+  },
+  "postCreateCommand": "",
+  "features": { "ghcr.io/devcontainers/features/git:1": {} },
+  "remoteUser": "vscode"
 }

.devcontainer/docker-compose.yml

@@ -1,17 +1,9 @@
-version: '3.4'
-
 services:
   app:
-    # container_name: LibreChat_dev
-    image: node:19-bullseye
-    # Using a Dockerfile is optional, but included for completeness.
-    # build: 
-    #   context: .
-    #   dockerfile: Dockerfile
-    #   # [Optional] You can use build args to set options. e.g. 'VARIANT' below affects the image in the Dockerfile
-    #   args: 
-    #     VARIANT: buster
-    # network_mode: "host"
+    build:
+      context: ..
+      dockerfile: .devcontainer/Dockerfile
+    # restart: always
     links:
       - mongodb
       - meilisearch
@@ -21,17 +13,16 @@ services:
     - "host.docker.internal:host-gateway"
 
     volumes:
-    #   # This is where VS Code should expect to find your project's source code and the value of "workspaceFolder" in .devcontainer/devcontainer.json
-      - ..:/workspace:cached
-    #   # - /app/client/node_modules
-    #   # - ./api:/app/api
-    #   # - ./.env:/app/.env
-    #   # - ./.env.development:/app/.env.development
-    #   # - ./.env.production:/app/.env.production
-    #   # - /app/api/node_modules
-
-    #   # Uncomment the next line to use Docker from inside the container. See https://aka.ms/vscode-remote/samples/docker-from-docker-compose for details.
-    #   # - /var/run/docker.sock:/var/run/docker.sock 
+      # This is where VS Code should expect to find your project's source code and the value of "workspaceFolder" in .devcontainer/devcontainer.json
+      - ..:/workspaces:cached
+      # Uncomment the next line to use Docker from inside the container. See https://aka.ms/vscode-remote/samples/docker-from-docker-compose for details.
+      # - /var/run/docker.sock:/var/run/docker.sock 
+    environment:
+      - HOST=0.0.0.0
+      - MONGO_URI=mongodb://mongodb:27017/LibreChat
+      # - CHATGPT_REVERSE_PROXY=http://host.docker.internal:8080/api/conversation # if you are hosting your own chatgpt reverse proxy with docker
+      # - OPENAI_REVERSE_PROXY=http://host.docker.internal:8070/v1/chat/completions # if you are hosting your own chatgpt reverse proxy with docker
+      - MEILI_HOST=http://meilisearch:7700
 
     # Runs app on the same network as the service container, allows "forwardPorts" in devcontainer.json function.
     # network_mode: service:another-service
@@ -39,45 +30,34 @@ services:
     # Use "forwardPorts" in **devcontainer.json** to forward an app port locally. 
     # (Adding the "ports" property to this file will not forward from a Codespace.)
 
-    # Uncomment the next line to use a non-root user for all processes - See https://aka.ms/vscode-remote/containers/non-root for details.
-    # user: vscode
-
-    # Uncomment the next four lines if you will use a ptrace-based debugger like C++, Go, and Rust.
-    # cap_add:
-    #   - SYS_PTRACE
-    # security_opt:
-    #   - seccomp:unconfined
+    # Use a non-root user for all processes - See https://aka.ms/vscode-remote/containers/non-root for details.
+    user: vscode
 
     # Overrides default command so things don't shut down after the process ends.
     command: /bin/sh -c "while sleep 1000; do :; done"  
 
   mongodb:
     container_name: chat-mongodb
-    # network_mode: "host"
     expose:
       - 27017
     # ports:
     #   - 27018:27017
     image: mongo
-  #   restart: always
+    # restart: always
     volumes:
       - ./data-node:/data/db
     command: mongod --noauth
   meilisearch:
     container_name: chat-meilisearch
-    image: getmeili/meilisearch:v1.0
-    # network_mode: "host"
+    image: getmeili/meilisearch:v1.5
+    # restart: always
     expose:
       - 7700
-    # ports:
-    #   - 7700:7700
-    # env_file:
-    #   - .env
+    # Uncomment this to access meilisearch from outside docker
+    # ports: 
+    #   - 7700:7700 # if exposing these ports, make sure your master key is not the default value
     environment:
-      - SEARCH=false
-      - MEILI_HOST=http://0.0.0.0:7700
-      - MEILI_HTTP_ADDR=0.0.0.0:7700
+      - MEILI_NO_ANALYTICS=true
       - MEILI_MASTER_KEY=5c71cf56d672d009e36070b5bc5e47b743535ae55c818ae3b735bb6ebfb4ba63
     volumes:
-      - ./meili_data:/meili_data
-  
+      - ./meili_data_v1.5:/meili_data

.dockerignore

@@ -1,5 +1,17 @@
+**/.circleci
+**/.editorconfig
+**/.dockerignore
+**/.git
+**/.DS_Store
+**/.vscode
 **/node_modules
-client/dist/images
+
+# Specific patterns to ignore
 data-node
-.env
-**/.env
+meili_data*
+librechat*
+Dockerfile*
+docs
+
+# Ignore all hidden files
+.*

.env.example

@@ -1,48 +1,333 @@
-##########################
-# Server configuration: 
-##########################
+#=====================================================================#
+#                       LibreChat Configuration                       #
+#=====================================================================#
+# Please refer to the reference documentation for assistance          #
+# with configuring your LibreChat environment.                        #
+#                                                                     #
+# https://www.librechat.ai/docs/configuration/dotenv                  #
+#=====================================================================#
 
-APP_TITLE=LibreChat
+#==================================================#
+#               Server Configuration               #
+#==================================================#
 
-# Uncomment to add a custom footer. 
-# Uncomment and make empty "" to remove the footer.
-# CUSTOM_FOOTER="My custom footer"
-
-# The server will listen to localhost:3080 by default. You can change the target IP as you want. 
-# If you want to make this server available externally, for example to share the server with others 
-# or expose this from a Docker container, set host to 0.0.0.0 or your external IP interface. 
-# Tips: Setting host to 0.0.0.0 means listening on all interfaces. It's not a real IP.  
-# Use localhost:port rather than 0.0.0.0:port to access the server. 
-# Set Node env to development if running in dev mode.
 HOST=localhost
 PORT=3080
 
-# Note: the following enables user balances, which you can add manually
-# or you will need to build out a balance accruing system for users.
-# For more info, see https://docs.librechat.ai/features/token_usage.html
+MONGO_URI=mongodb://127.0.0.1:27017/LibreChat
 
-# To manually add balances, run the following command:
-# `npm run add-balance`
+DOMAIN_CLIENT=http://localhost:3080
+DOMAIN_SERVER=http://localhost:3080
 
-# You can also specify the email and token credit amount to add, e.g.:
-# `npm run add-balance example@example.com 1000`
+NO_INDEX=true
+# Use the address that is at most n number of hops away from the Express application. 
+# req.socket.remoteAddress is the first hop, and the rest are looked for in the X-Forwarded-For header from right to left. 
+# A value of 0 means that the first untrusted address would be req.socket.remoteAddress, i.e. there is no reverse proxy.
+# Defaulted to 1.
+TRUST_PROXY=1
 
-# This works well to track your own usage for personal use; 1000 credits = $0.001 (1 mill USD)
+#===============#
+# JSON Logging  #
+#===============#
 
-# Set to true to enable token credit balances for the OpenAI/Plugins endpoints
-CHECK_BALANCE=false
+# Use when process console logs in cloud deployment like GCP/AWS
+CONSOLE_JSON=false
 
-# Automated Moderation System
-# The Automated Moderation System uses a scoring mechanism to track user violations. As users commit actions
-# like excessive logins, registrations, or messaging, they accumulate violation scores. Upon reaching
-# a set threshold, the user and their IP are temporarily banned. This system ensures platform security
-# by monitoring and penalizing rapid or suspicious activities.
+#===============#
+# Debug Logging #
+#===============#
 
-BAN_VIOLATIONS=true # Whether or not to enable banning users for violations (they will still be logged)
-BAN_DURATION=1000 * 60 * 60 * 2 # how long the user and associated IP are banned for
-BAN_INTERVAL=20 # a user will be banned everytime their score reaches/crosses over the interval threshold
+DEBUG_LOGGING=true
+DEBUG_CONSOLE=false
 
-# The score for each violation
+#=============#
+# Permissions #
+#=============#
+
+# UID=1000
+# GID=1000
+
+#===============#
+# Configuration #
+#===============#
+# Use an absolute path, a relative path, or a URL
+
+# CONFIG_PATH="/alternative/path/to/librechat.yaml"
+
+#===================================================#
+#                     Endpoints                     #
+#===================================================#
+
+# ENDPOINTS=openAI,assistants,azureOpenAI,google,gptPlugins,anthropic
+
+PROXY=
+
+#===================================#
+# Known Endpoints - librechat.yaml  #
+#===================================#
+# https://www.librechat.ai/docs/configuration/librechat_yaml/ai_endpoints
+
+# ANYSCALE_API_KEY=
+# APIPIE_API_KEY=
+# COHERE_API_KEY=
+# DEEPSEEK_API_KEY=
+# DATABRICKS_API_KEY=
+# FIREWORKS_API_KEY=
+# GROQ_API_KEY=
+# HUGGINGFACE_TOKEN=
+# MISTRAL_API_KEY=
+# OPENROUTER_KEY=
+# PERPLEXITY_API_KEY=
+# SHUTTLEAI_API_KEY=
+# TOGETHERAI_API_KEY=
+# UNIFY_API_KEY=
+# XAI_API_KEY=
+
+#============#
+# Anthropic  #
+#============#
+
+ANTHROPIC_API_KEY=user_provided
+# ANTHROPIC_MODELS=claude-3-7-sonnet-latest,claude-3-7-sonnet-20250219,claude-3-5-haiku-20241022,claude-3-5-sonnet-20241022,claude-3-5-sonnet-latest,claude-3-5-sonnet-20240620,claude-3-opus-20240229,claude-3-sonnet-20240229,claude-3-haiku-20240307,claude-2.1,claude-2,claude-1.2,claude-1,claude-1-100k,claude-instant-1,claude-instant-1-100k
+# ANTHROPIC_REVERSE_PROXY=
+
+#============#
+# Azure      #
+#============#
+
+# Note: these variables are DEPRECATED
+# Use the `librechat.yaml` configuration for `azureOpenAI` instead
+# You may also continue to use them if you opt out of using the `librechat.yaml` configuration
+
+# AZURE_OPENAI_DEFAULT_MODEL=gpt-3.5-turbo # Deprecated
+# AZURE_OPENAI_MODELS=gpt-3.5-turbo,gpt-4 # Deprecated
+# AZURE_USE_MODEL_AS_DEPLOYMENT_NAME=TRUE # Deprecated
+# AZURE_API_KEY= # Deprecated
+# AZURE_OPENAI_API_INSTANCE_NAME= # Deprecated
+# AZURE_OPENAI_API_DEPLOYMENT_NAME= # Deprecated
+# AZURE_OPENAI_API_VERSION= # Deprecated
+# AZURE_OPENAI_API_COMPLETIONS_DEPLOYMENT_NAME= # Deprecated
+# AZURE_OPENAI_API_EMBEDDINGS_DEPLOYMENT_NAME= # Deprecated
+# PLUGINS_USE_AZURE="true" # Deprecated
+
+#=================#
+#   AWS Bedrock   #
+#=================#
+
+# BEDROCK_AWS_DEFAULT_REGION=us-east-1 # A default region must be provided
+# BEDROCK_AWS_ACCESS_KEY_ID=someAccessKey
+# BEDROCK_AWS_SECRET_ACCESS_KEY=someSecretAccessKey
+# BEDROCK_AWS_SESSION_TOKEN=someSessionToken
+
+# Note: This example list is not meant to be exhaustive. If omitted, all known, supported model IDs will be included for you.
+# BEDROCK_AWS_MODELS=anthropic.claude-3-5-sonnet-20240620-v1:0,meta.llama3-1-8b-instruct-v1:0
+
+# See all Bedrock model IDs here: https://docs.aws.amazon.com/bedrock/latest/userguide/model-ids.html#model-ids-arns
+
+# Notes on specific models:
+# The following models are not support due to not supporting streaming:
+# ai21.j2-mid-v1
+
+# The following models are not support due to not supporting conversation history:
+# ai21.j2-ultra-v1, cohere.command-text-v14, cohere.command-light-text-v14
+
+#============#
+# Google     #
+#============#
+
+GOOGLE_KEY=user_provided
+
+# GOOGLE_REVERSE_PROXY=
+# Some reverse proxies do not support the X-goog-api-key header, uncomment to pass the API key in Authorization header instead.
+# GOOGLE_AUTH_HEADER=true
+
+# Gemini API (AI Studio)
+# GOOGLE_MODELS=gemini-2.0-flash-exp,gemini-2.0-flash-thinking-exp-1219,gemini-exp-1121,gemini-exp-1114,gemini-1.5-flash-latest,gemini-1.0-pro,gemini-1.0-pro-001,gemini-1.0-pro-latest,gemini-1.0-pro-vision-latest,gemini-1.5-pro-latest,gemini-pro,gemini-pro-vision
+
+# Vertex AI
+# GOOGLE_MODELS=gemini-1.5-flash-preview-0514,gemini-1.5-pro-preview-0514,gemini-1.0-pro-vision-001,gemini-1.0-pro-002,gemini-1.0-pro-001,gemini-pro-vision,gemini-1.0-pro
+
+# GOOGLE_TITLE_MODEL=gemini-pro
+
+# GOOGLE_LOC=us-central1
+
+# Google Safety Settings
+# NOTE: These settings apply to both Vertex AI and Gemini API (AI Studio)
+#
+# For Vertex AI:
+# To use the BLOCK_NONE setting, you need either:
+# (a) Access through an allowlist via your Google account team, or
+# (b) Switch to monthly invoiced billing: https://cloud.google.com/billing/docs/how-to/invoiced-billing
+#
+# For Gemini API (AI Studio):
+# BLOCK_NONE is available by default, no special account requirements.
+#
+# Available options: BLOCK_NONE, BLOCK_ONLY_HIGH, BLOCK_MEDIUM_AND_ABOVE, BLOCK_LOW_AND_ABOVE
+#
+# GOOGLE_SAFETY_SEXUALLY_EXPLICIT=BLOCK_ONLY_HIGH
+# GOOGLE_SAFETY_HATE_SPEECH=BLOCK_ONLY_HIGH
+# GOOGLE_SAFETY_HARASSMENT=BLOCK_ONLY_HIGH
+# GOOGLE_SAFETY_DANGEROUS_CONTENT=BLOCK_ONLY_HIGH
+# GOOGLE_SAFETY_CIVIC_INTEGRITY=BLOCK_ONLY_HIGH
+
+#============#
+# OpenAI     #
+#============#
+
+OPENAI_API_KEY=user_provided
+# OPENAI_MODELS=o1,o1-mini,o1-preview,gpt-4o,chatgpt-4o-latest,gpt-4o-mini,gpt-3.5-turbo-0125,gpt-3.5-turbo-0301,gpt-3.5-turbo,gpt-4,gpt-4-0613,gpt-4-vision-preview,gpt-3.5-turbo-0613,gpt-3.5-turbo-16k-0613,gpt-4-0125-preview,gpt-4-turbo-preview,gpt-4-1106-preview,gpt-3.5-turbo-1106,gpt-3.5-turbo-instruct,gpt-3.5-turbo-instruct-0914,gpt-3.5-turbo-16k
+
+DEBUG_OPENAI=false
+
+# TITLE_CONVO=false
+# OPENAI_TITLE_MODEL=gpt-4o-mini
+
+# OPENAI_SUMMARIZE=true
+# OPENAI_SUMMARY_MODEL=gpt-4o-mini
+
+# OPENAI_FORCE_PROMPT=true
+
+# OPENAI_REVERSE_PROXY=
+
+# OPENAI_ORGANIZATION=
+
+#====================#
+#   Assistants API   #
+#====================#
+
+ASSISTANTS_API_KEY=user_provided
+# ASSISTANTS_BASE_URL=
+# ASSISTANTS_MODELS=gpt-4o,gpt-4o-mini,gpt-3.5-turbo-0125,gpt-3.5-turbo-16k-0613,gpt-3.5-turbo-16k,gpt-3.5-turbo,gpt-4,gpt-4-0314,gpt-4-32k-0314,gpt-4-0613,gpt-3.5-turbo-0613,gpt-3.5-turbo-1106,gpt-4-0125-preview,gpt-4-turbo-preview,gpt-4-1106-preview
+
+#==========================#
+#   Azure Assistants API   #
+#==========================#
+
+# Note: You should map your credentials with custom variables according to your Azure OpenAI Configuration
+# The models for Azure Assistants are also determined by your Azure OpenAI configuration.
+
+# More info, including how to enable use of Assistants with Azure here:
+# https://www.librechat.ai/docs/configuration/librechat_yaml/ai_endpoints/azure#using-assistants-with-azure
+
+#============#
+# Plugins    #
+#============#
+
+# PLUGIN_MODELS=gpt-4o,gpt-4o-mini,gpt-4,gpt-4-turbo-preview,gpt-4-0125-preview,gpt-4-1106-preview,gpt-4-0613,gpt-3.5-turbo,gpt-3.5-turbo-0125,gpt-3.5-turbo-1106,gpt-3.5-turbo-0613
+
+DEBUG_PLUGINS=true
+
+CREDS_KEY=f34be427ebb29de8d88c107a71546019685ed8b241d8f2ed00c3df97ad2566f0
+CREDS_IV=e2341419ec3dd3d19b13a1a87fafcbfb
+
+# Azure AI Search
+#-----------------
+AZURE_AI_SEARCH_SERVICE_ENDPOINT=
+AZURE_AI_SEARCH_INDEX_NAME=
+AZURE_AI_SEARCH_API_KEY=
+
+AZURE_AI_SEARCH_API_VERSION=
+AZURE_AI_SEARCH_SEARCH_OPTION_QUERY_TYPE=
+AZURE_AI_SEARCH_SEARCH_OPTION_TOP=
+AZURE_AI_SEARCH_SEARCH_OPTION_SELECT=
+
+# DALL·E
+#----------------
+# DALLE_API_KEY=
+# DALLE3_API_KEY=
+# DALLE2_API_KEY=
+# DALLE3_SYSTEM_PROMPT=
+# DALLE2_SYSTEM_PROMPT=
+# DALLE_REVERSE_PROXY=
+# DALLE3_BASEURL=
+# DALLE2_BASEURL=
+
+# DALL·E (via Azure OpenAI)
+# Note: requires some of the variables above to be set
+#----------------
+# DALLE3_AZURE_API_VERSION=
+# DALLE2_AZURE_API_VERSION=
+
+
+# Google
+#-----------------
+GOOGLE_SEARCH_API_KEY=
+GOOGLE_CSE_ID=
+
+# YOUTUBE
+#-----------------
+YOUTUBE_API_KEY=
+
+# SerpAPI
+#-----------------
+SERPAPI_API_KEY=
+
+# Stable Diffusion
+#-----------------
+SD_WEBUI_URL=http://host.docker.internal:7860
+
+# Tavily
+#-----------------
+TAVILY_API_KEY=
+
+# Traversaal
+#-----------------
+TRAVERSAAL_API_KEY=
+
+# WolframAlpha
+#-----------------
+WOLFRAM_APP_ID=
+
+# Zapier
+#-----------------
+ZAPIER_NLA_API_KEY=
+
+#==================================================#
+#                      Search                      #
+#==================================================#
+
+SEARCH=true
+MEILI_NO_ANALYTICS=true
+MEILI_HOST=http://0.0.0.0:7700
+MEILI_MASTER_KEY=DrhYf7zENyR6AlUCKmnz0eYASOQdl6zxH7s7MKFSfFCt
+
+# Optional: Disable indexing, useful in a multi-node setup
+# where only one instance should perform an index sync.
+# MEILI_NO_SYNC=true
+
+#==================================================#
+#          Speech to Text & Text to Speech         #
+#==================================================#
+
+STT_API_KEY=
+TTS_API_KEY=
+
+#==================================================#
+#                        RAG                       #
+#==================================================#
+# More info: https://www.librechat.ai/docs/configuration/rag_api
+
+# RAG_OPENAI_BASEURL=
+# RAG_OPENAI_API_KEY=
+# RAG_USE_FULL_CONTEXT=
+# EMBEDDINGS_PROVIDER=openai
+# EMBEDDINGS_MODEL=text-embedding-3-small
+
+#===================================================#
+#                    User System                    #
+#===================================================#
+
+#========================#
+# Moderation             #
+#========================#
+
+OPENAI_MODERATION=false
+OPENAI_MODERATION_API_KEY=
+# OPENAI_MODERATION_REVERSE_PROXY=
+
+BAN_VIOLATIONS=true
+BAN_DURATION=1000 * 60 * 60 * 2
+BAN_INTERVAL=20
 
 LOGIN_VIOLATION_SCORE=1
 REGISTRATION_VIOLATION_SCORE=1
@@ -50,398 +335,206 @@ CONCURRENT_VIOLATION_SCORE=1
 MESSAGE_VIOLATION_SCORE=1
 NON_BROWSER_VIOLATION_SCORE=20
 
-# Login and registration rate limiting.
+LOGIN_MAX=7
+LOGIN_WINDOW=5
+REGISTER_MAX=5
+REGISTER_WINDOW=60
 
-LOGIN_MAX=7 # The max amount of logins allowed per IP per LOGIN_WINDOW
-LOGIN_WINDOW=5 # in minutes, determines the window of time for LOGIN_MAX logins
-REGISTER_MAX=5 # The max amount of registrations allowed per IP per REGISTER_WINDOW
-REGISTER_WINDOW=60 # in minutes, determines the window of time for REGISTER_MAX registrations
+LIMIT_CONCURRENT_MESSAGES=true
+CONCURRENT_MESSAGE_MAX=2
 
-# Message rate limiting (per user & IP)
+LIMIT_MESSAGE_IP=true
+MESSAGE_IP_MAX=40
+MESSAGE_IP_WINDOW=1
 
-LIMIT_CONCURRENT_MESSAGES=true # Whether to limit the amount of messages a user can send per request
-CONCURRENT_MESSAGE_MAX=2 # The max amount of messages a user can send per request
+LIMIT_MESSAGE_USER=false
+MESSAGE_USER_MAX=40
+MESSAGE_USER_WINDOW=1
 
-LIMIT_MESSAGE_IP=true # Whether to limit the amount of messages an IP can send per MESSAGE_IP_WINDOW
-MESSAGE_IP_MAX=40 # The max amount of messages an IP can send per MESSAGE_IP_WINDOW
-MESSAGE_IP_WINDOW=1 # in minutes, determines the window of time for MESSAGE_IP_MAX messages
+ILLEGAL_MODEL_REQ_SCORE=5
 
-# Note: You can utilize both limiters, but default is to limit by IP only.
-LIMIT_MESSAGE_USER=false # Whether to limit the amount of messages an IP can send per MESSAGE_USER_WINDOW
-MESSAGE_USER_MAX=40 # The max amount of messages an IP can send per MESSAGE_USER_WINDOW
-MESSAGE_USER_WINDOW=1 # in minutes, determines the window of time for MESSAGE_USER_MAX messages
+#========================#
+# Balance                #
+#========================#
 
-# If you have permission problems, set here the UID and GID of the user running
-# the docker compose command. The applications in the container will run with these uid/gid.
-# UID=1000
-# GID=1000
+CHECK_BALANCE=false
+# START_BALANCE=20000 # note: the number of tokens that will be credited after registration.
 
-# Change this to proxy any API request. 
-# It's useful if your machine has difficulty calling the original API server. 
-# PROXY=
+#========================#
+# Registration and Login #
+#========================#
 
-# Change this to your MongoDB URI if different. I recommend appending LibreChat. 
-MONGO_URI=mongodb://127.0.0.1:27018/LibreChat
-
-##########################
-# OpenAI Endpoint: 
-##########################
-
-# Access key from OpenAI platform. 
-# Leave it blank to disable this feature.
-# Set to "user_provided" to allow the user to provide their API key from the UI.
-OPENAI_API_KEY=user_provided
-
-DEBUG_OPENAI=false # Set to true to enable debug mode for the OpenAI endpoint
-
-# Identify the available models, separated by commas *without spaces*. 
-# The first will be default. 
-# Leave it blank to use internal settings. 
-# OPENAI_MODELS=gpt-3.5-turbo-1106,gpt-4-1106-preview,gpt-3.5-turbo,gpt-3.5-turbo-16k,gpt-3.5-turbo-0301,text-davinci-003,gpt-4,gpt-4-0314,gpt-4-0613
-
-# Titling is enabled by default when initiating a conversation.
-# Uncomment the following variable to disable this feature.
-# TITLE_CONVO=false
-
-# (Optional) The default model used for titling by is gpt-3.5-turbo-0613
-# You can change it by uncommenting the following and setting the desired model
-# Must be compatible with the OpenAI Endpoint.
-# OPENAI_TITLE_MODEL=gpt-3.5-turbo
-
-# (Optional/Experimental) Enable message summarization by uncommenting the following:
-# Note: this may affect response time when a summary is being generated.
-# OPENAI_SUMMARIZE=true
-
-# Not yet implemented: this will be a conversation option enabled by default to save users on tokens
-# We are using the ConversationSummaryBufferMemory method to summarize messages.
-# To learn more about this, see this article:
-# https://www.pinecone.io/learn/series/langchain/langchain-conversational-memory/
-
-# (Optional) The default model used for summarizing is gpt-3.5-turbo
-# You can change it by uncommenting the following and setting the desired model
-# Must be compatible with the OpenAI Endpoint.
-# OPENAI_SUMMARY_MODEL=gpt-3.5-turbo
-
-# Reverse proxy settings for OpenAI: 
-# https://github.com/waylaidwanderer/node-chatgpt-api#using-a-reverse-proxy
-# The URL must match the "url/v1," pattern, the "openai" suffix is also allowed.
-# Examples:
-#   - https://open.ai/v1
-#   - https://open.ai/v1/ACCOUNT/GATEWAY/openai
-#   - https://open.ai/v1/hi/openai
-
-# OPENAI_REVERSE_PROXY=
-
-# (Advanced) Sometimes when using Local LLM APIs, you may need to force the API
-# to be called with a `prompt` payload instead of a `messages` payload; to mimic the
-# a `/v1/completions` request instead of `/v1/chat/completions`
-# This may be the case for LocalAI with some models. To do so, uncomment the following:
-# OPENAI_FORCE_PROMPT=true
-
-# (Advanced) For customization of the DALL-E-3 System prompt, 
-# uncomment the following, and provide your own prompt:
-# See official prompt for reference:
-# https://github.com/spdustin/ChatGPT-AutoExpert/blob/main/_system-prompts/dall-e.md
-# DALLE3_SYSTEM_PROMPT="Your System Prompt here"
-
-# (Advanced) DALL-E Proxy settings
-# This is separate from its OpenAI counterpart for customization purposes
-
-# Reverse proxy settings, changes the baseURL for the DALL-E-3 API Calls
-# The URL must match the "url/v1," pattern, the "openai" suffix is also allowed.
-# Examples:
-#   - https://open.ai/v1
-#   - https://open.ai/v1/ACCOUNT/GATEWAY/openai
-#   - https://open.ai/v1/hi/openai
-
-# DALLE_REVERSE_PROXY=
-
-# Note: if you have PROXY set, it will be used for DALLE calls also, which is universal for the app
-
-##########################
-# OpenRouter (overrides OpenAI and Plugins Endpoints): 
-##########################
-
-# OpenRouter is a legitimate proxy service to a multitude of LLMs, both closed and open source, including:
-# OpenAI models, Anthropic models, Meta's Llama models, pygmalionai/mythalion-13b
-# and many more open source models. Newer integrations are usually discounted, too!
-
-# Note: this overrides the OpenAI and Plugins Endpoints.
-# See ./docs/install/free_ai_apis.md for more info.
-
-# OPENROUTER_API_KEY=
-
-##########################
-# AZURE Endpoint:   
-##########################
-
-# To use Azure with this project, set the following variables. These will be used to build the API URL.
-# Chat completion: 
-# `https://{AZURE_OPENAI_API_INSTANCE_NAME}.openai.azure.com/openai/deployments/{AZURE_OPENAI_API_DEPLOYMENT_NAME}/chat/completions?api-version={AZURE_OPENAI_API_VERSION}`;
-# You should also consider changing the `OPENAI_MODELS` variable above to the models available in your instance/deployment.
-# Note: I've noticed that the Azure API is much faster than the OpenAI API, so the streaming looks almost instantaneous.
-# Note "AZURE_OPENAI_API_COMPLETIONS_DEPLOYMENT_NAME" and "AZURE_OPENAI_API_EMBEDDINGS_DEPLOYMENT_NAME" are optional but might be used in the future
-
-# AZURE_API_KEY=
-# AZURE_OPENAI_API_INSTANCE_NAME=
-# AZURE_OPENAI_API_DEPLOYMENT_NAME=
-# AZURE_OPENAI_API_VERSION=
-# AZURE_OPENAI_API_COMPLETIONS_DEPLOYMENT_NAME=
-# AZURE_OPENAI_API_EMBEDDINGS_DEPLOYMENT_NAME=
-
-# NOTE: As of 2023-11-10, the Azure API only allows one model per deployment,
-# It's recommended to name your deployments after the model name, e.g. "gpt-35-turbo,"
-# which allows for fast deployment switching and AZURE_USE_MODEL_AS_DEPLOYMENT_NAME enabled.
-# However, you can use non-model deployment names and setting the AZURE_OPENAI_DEFAULT_MODEL to ensure it works as expected.
-
-# Identify the available models, separated by commas *without spaces*.
-# The first will be default. Leave it blank or as is to use internal settings. 
-# NOTE: as deployment names can't have periods, they will be removed when the endpoint is generated.
-AZURE_OPENAI_MODELS=gpt-3.5-turbo,gpt-4
-
-# (Advanced) this enables the use of the model name as the deployment name, e.g. "gpt-3.5-turbo" as the deployment name
-AZURE_USE_MODEL_AS_DEPLOYMENT_NAME=TRUE
-
-# (Advanced) this overrides the model setting for Azure, in case you want to use your custom deployment names
-# as the values for AZURE_OPENAI_MODELS
-# AZURE_OPENAI_DEFAULT_MODEL=gpt-3.5-turbo
-
-# To use Azure with the Plugins endpoint, you need the variables above, and uncomment the following variable:
-# NOTE: This may not work as expected and Azure OpenAI may not support OpenAI Functions yet
-# Omit/leave it commented to use the default OpenAI API
-
-# PLUGINS_USE_AZURE="true"
-
-##########################
-# ChatGPT Endpoint: 
-##########################
-
-# ChatGPT Browser Client (free but use at your own risk) 
-# Access token from https://chat.openai.com/api/auth/session 
-# Exposes your access token to `CHATGPT_REVERSE_PROXY`
-# Set to "user_provided" to allow the user to provide its token from the UI.
-# Leave it blank to disable this endpoint 
-CHATGPT_TOKEN=user_provided
-
-# Identify the available models, separated by commas. The first will be default. 
-# Leave it blank to use internal settings. 
-CHATGPT_MODELS=text-davinci-002-render-sha,gpt-4
-# NOTE: you can add gpt-4-plugins, gpt-4-code-interpreter, and gpt-4-browsing to the list above and use the models for these features;
-# however, the view/display portion of these features are not supported, but you can use the underlying models, which have higher token context
-# Also: text-davinci-002-render-paid is deprecated as of May 2023
-
-# Reverse proxy setting for OpenAI
-# https://github.com/waylaidwanderer/node-chatgpt-api#using-a-reverse-proxy
-# By default it will use the node-chatgpt-api recommended proxy, (it's a third party server)
-# CHATGPT_REVERSE_PROXY=<YOUR REVERSE PROXY>
-
-##########################
-# BingAI Endpoint:   
-##########################
-
-# Also used for Sydney and jailbreak
-# To get your Access token for Bing, login to https://www.bing.com
-# Use dev tools or an extension while logged into the site to copy the content of the _U cookie. 
-# If this fails, follow these instructions https://github.com/danny-avila/LibreChat/issues/370#issuecomment-1560382302 to provide the full cookie strings 
-# or check out our discord https://discord.com/channels/1086345563026489514/1143941308684177429
-# Set to "user_provided" to allow the user to provide its token from the UI.
-# Leave it blank to disable this endpoint. 
-BINGAI_TOKEN=user_provided
-
-# BingAI Host: 
-# Necessary for some people in different countries, e.g. China (https://cn.bing.com) 
-# Leave it blank to use default server. 
-# BINGAI_HOST=https://cn.bing.com 
-
-#############################
-# Plugins:
-#############################
-
-# Identify the available models, separated by commas *without spaces*. 
-# The first will be default. 
-# Leave it blank to use internal settings. 
-# PLUGIN_MODELS=gpt-3.5-turbo,gpt-3.5-turbo-16k,gpt-3.5-turbo-0301,gpt-4,gpt-4-0314,gpt-4-0613
-
-DEBUG_PLUGINS=true # Set to false or comment out to disable debug mode for plugins
-
-# For securely storing credentials, you need a fixed key and IV. You can set them here for prod and dev environments
-# If you don't set them, the app will crash on startup.
-# You need a 32-byte key (64 characters in hex) and 16-byte IV (32 characters in hex)
-# Use this replit to generate some quickly: https://replit.com/@daavila/crypto#index.js
-# Here are some examples (THESE ARE NOT SECURE!)
-CREDS_KEY=f34be427ebb29de8d88c107a71546019685ed8b241d8f2ed00c3df97ad2566f0
-CREDS_IV=e2341419ec3dd3d19b13a1a87fafcbfb
-
-# AI-Assisted Google Search  
-# This bot supports searching google for answers to your questions with assistance from GPT! 
-# See detailed instructions here: https://github.com/danny-avila/LibreChat/blob/main/docs/features/plugins/google_search.md
-GOOGLE_API_KEY=
-GOOGLE_CSE_ID=
-
-# StableDiffusion WebUI
-# This bot supports StableDiffusion WebUI, using it's API to generated requested images.
-# See detailed instructions here: https://github.com/danny-avila/LibreChat/blob/main/docs/features/plugins/stable_diffusion.md
-# Use "http://127.0.0.1:7860" with local install and "http://host.docker.internal:7860" for docker
-SD_WEBUI_URL=http://host.docker.internal:7860
-
-# Azure Cognitive Search
-# This plugin supports searching Azure Cognitive Search for answers to your questions.
-# See detailed instructions here: https://github.com/danny-avila/LibreChat/blob/main/docs/features/plugins/azure_cognitive_search.md
-AZURE_COGNITIVE_SEARCH_SERVICE_ENDPOINT=
-AZURE_COGNITIVE_SEARCH_INDEX_NAME=
-AZURE_COGNITIVE_SEARCH_API_KEY=
-
-AZURE_COGNITIVE_SEARCH_API_VERSION=
-AZURE_COGNITIVE_SEARCH_SEARCH_OPTION_QUERY_TYPE=
-AZURE_COGNITIVE_SEARCH_SEARCH_OPTION_TOP=
-AZURE_COGNITIVE_SEARCH_SEARCH_OPTION_SELECT=
-
-##########################
-# PaLM (Google) Endpoint: 
-##########################
-
-# Follow the instruction here to setup:
-# https://github.com/danny-avila/LibreChat/blob/main/docs/install/apis_and_tokens.md
-
-PALM_KEY=user_provided
-
-# In case you need a reverse proxy for this endpoint:
-# GOOGLE_REVERSE_PROXY= 
-
-##########################
-# Anthropic Endpoint: 
-##########################
-# Access key from https://console.anthropic.com/
-# Leave it blank to disable this feature.
-# Set to "user_provided" to allow the user to provide their API key from the UI.
-# Note that access to claude-1 may potentially become unavailable with the release of claude-2.
-ANTHROPIC_API_KEY=user_provided
-ANTHROPIC_MODELS=claude-1,claude-instant-1,claude-2
-
-##########################
-# Proxy: To be Used by all endpoints
-##########################
-
-PROXY=
-
-##########################
-# Search: 
-##########################
-
-# ENABLING SEARCH MESSAGES/CONVOS 
-# Requires the installation of the free self-hosted Meilisearch or a paid Remote Plan (Remote not tested) 
-# The easiest setup for this is through docker-compose, which takes care of it for you. 
-SEARCH=true
-
-# HIGHLY RECOMMENDED: Disable anonymized telemetry analytics for MeiliSearch for absolute privacy.
-MEILI_NO_ANALYTICS=true
-
-# REQUIRED FOR SEARCH: MeiliSearch Host, mainly for the API server to connect to the search server. 
-# Replace '0.0.0.0' with 'meilisearch' if serving MeiliSearch with docker-compose. 
-MEILI_HOST=http://0.0.0.0:7700
-
-# REQUIRED FOR SEARCH: MeiliSearch HTTP Address, mainly for docker-compose to expose the search server. 
-# Replace '0.0.0.0' with 'meilisearch' if serving MeiliSearch with docker-compose. 
-MEILI_HTTP_ADDR=0.0.0.0:7700
-
-# REQUIRED FOR SEARCH: In production env., a secure key is needed. You can generate your own. 
-# This master key must be at least 16 bytes, composed of valid UTF-8 characters. 
-# MeiliSearch will throw an error and refuse to launch if no master key is provided, 
-# or if it is under 16 bytes. MeiliSearch will suggest a secure autogenerated master key.
-# Using docker, it seems recognized as production so use a secure key.
-# This is a ready made secure key for docker-compose, you can replace it with your own.
-MEILI_MASTER_KEY=DrhYf7zENyR6AlUCKmnz0eYASOQdl6zxH7s7MKFSfFCt
-
-##########################
-# User System:
-##########################
-
-# Allow Public Registration
+ALLOW_EMAIL_LOGIN=true
 ALLOW_REGISTRATION=true
-
-# Allow Social Registration
 ALLOW_SOCIAL_LOGIN=false
-
-# Allow Social Registration (WORKS ONLY for Google, Github, Discord)
 ALLOW_SOCIAL_REGISTRATION=false
+ALLOW_PASSWORD_RESET=false
+# ALLOW_ACCOUNT_DELETION=true # note: enabled by default if omitted/commented out
+ALLOW_UNVERIFIED_EMAIL_LOGIN=true
+
+SESSION_EXPIRY=1000 * 60 * 15
+REFRESH_TOKEN_EXPIRY=(1000 * 60 * 60 * 24) * 7
 
-# JWT Secrets
-# You should use secure values. The examples given are 32-byte keys (64 characters in hex)
-# Use this replit to generate some quickly: https://replit.com/@daavila/crypto#index.js
 JWT_SECRET=16f8c0ef4a5d391b26034086c628469d3f9f497f08163ab9b40137092f2909ef
 JWT_REFRESH_SECRET=eaa5191f2914e30b9387fd84e254e4ba6fc51b4654968a9b0803b456a54b8418
 
-# Google:
-# Add your Google Client ID and Secret here, you must register an app with Google Cloud to get these values
-# https://cloud.google.com/
-GOOGLE_CLIENT_ID=
-GOOGLE_CLIENT_SECRET=
-GOOGLE_CALLBACK_URL=/oauth/google/callback
+# Discord
+DISCORD_CLIENT_ID=
+DISCORD_CLIENT_SECRET=
+DISCORD_CALLBACK_URL=/oauth/discord/callback
 
-# Facebook:
-# Add your Facebook Client ID and Secret here, you must register an app with Facebook to get these values
-# https://developers.facebook.com/
+# Facebook
 FACEBOOK_CLIENT_ID=
 FACEBOOK_CLIENT_SECRET=
 FACEBOOK_CALLBACK_URL=/oauth/facebook/callback
 
-# OpenID:
-# See OpenID provider to get the below values
-# Create random string for OPENID_SESSION_SECRET
-# For Azure AD
-#   ISSUER:  https://login.microsoftonline.com/(tenant id)/v2.0/
-#   SCOPE:   openid profile email
+# GitHub
+GITHUB_CLIENT_ID=
+GITHUB_CLIENT_SECRET=
+GITHUB_CALLBACK_URL=/oauth/github/callback
+# GitHub Enterprise
+# GITHUB_ENTERPRISE_BASE_URL=
+# GITHUB_ENTERPRISE_USER_AGENT=
+
+# Google
+GOOGLE_CLIENT_ID=
+GOOGLE_CLIENT_SECRET=
+GOOGLE_CALLBACK_URL=/oauth/google/callback
+
+# Apple
+APPLE_CLIENT_ID=
+APPLE_TEAM_ID=
+APPLE_KEY_ID=
+APPLE_PRIVATE_KEY_PATH=
+APPLE_CALLBACK_URL=/oauth/apple/callback
+
+# OpenID
 OPENID_CLIENT_ID=
 OPENID_CLIENT_SECRET=
 OPENID_ISSUER=
 OPENID_SESSION_SECRET=
 OPENID_SCOPE="openid profile email"
 OPENID_CALLBACK_URL=/oauth/openid/callback
-# If LABEL and URL are left empty, then the default OpenID label and logo are used.
+OPENID_REQUIRED_ROLE=
+OPENID_REQUIRED_ROLE_TOKEN_KIND=
+OPENID_REQUIRED_ROLE_PARAMETER_PATH=
+# Set to determine which user info property returned from OpenID Provider to store as the User's username
+OPENID_USERNAME_CLAIM=
+# Set to determine which user info property returned from OpenID Provider to store as the User's name
+OPENID_NAME_CLAIM=
+
 OPENID_BUTTON_LABEL=
 OPENID_IMAGE_URL=
 
-# Set the expiration delay for the secure cookie with the JWT token
-# Recommend session expiry to be 15 minutes
-# Delay is in millisecond e.g. 7 days is 1000*60*60*24*7
-SESSION_EXPIRY=1000 * 60 * 15
-REFRESH_TOKEN_EXPIRY=(1000 * 60 * 60 * 24) * 7
+# LDAP
+LDAP_URL=
+LDAP_BIND_DN=
+LDAP_BIND_CREDENTIALS=
+LDAP_USER_SEARCH_BASE=
+LDAP_SEARCH_FILTER=mail={{username}}
+LDAP_CA_CERT_PATH=
+# LDAP_TLS_REJECT_UNAUTHORIZED=
+# LDAP_LOGIN_USES_USERNAME=true
+# LDAP_ID=
+# LDAP_USERNAME=
+# LDAP_EMAIL=
+# LDAP_FULL_NAME=
 
-# Github:
-# Get the Client ID and Secret from your Discord Application
-# Add your Discord Client ID and Client Secret here:
+#========================#
+# Email Password Reset   #
+#========================#
 
-GITHUB_CLIENT_ID=your_client_id
-GITHUB_CLIENT_SECRET=your_client_secret
-GITHUB_CALLBACK_URL=/oauth/github/callback # this should be the same for everyone
+EMAIL_SERVICE=
+EMAIL_HOST=
+EMAIL_PORT=25
+EMAIL_ENCRYPTION=
+EMAIL_ENCRYPTION_HOSTNAME=
+EMAIL_ALLOW_SELFSIGNED=
+EMAIL_USERNAME=
+EMAIL_PASSWORD=
+EMAIL_FROM_NAME=
+EMAIL_FROM=noreply@librechat.ai
 
-# Discord:
-# Get the Client ID and Secret from your Discord Application
-# Add your Github Client ID and Client Secret here:
+#========================#
+# Firebase CDN           #
+#========================#
 
-DISCORD_CLIENT_ID=your_client_id
-DISCORD_CLIENT_SECRET=your_client_secret
-DISCORD_CALLBACK_URL=/oauth/discord/callback # this should be the same for everyone
+FIREBASE_API_KEY=
+FIREBASE_AUTH_DOMAIN=
+FIREBASE_PROJECT_ID=
+FIREBASE_STORAGE_BUCKET=
+FIREBASE_MESSAGING_SENDER_ID=
+FIREBASE_APP_ID=
 
-###########################
-# Application Domains
-###########################
+#========================#
+# Shared Links           #
+#========================#
 
-# Note: 
-# Server = Backend 
-# Client = Public (the client is the url you visit)
-# For the Google login to work in dev mode, you will need to change DOMAIN_SERVER to localhost:3090 or place it in .env.development
+ALLOW_SHARED_LINKS=true
+ALLOW_SHARED_LINKS_PUBLIC=true
 
-DOMAIN_CLIENT=http://localhost:3080
-DOMAIN_SERVER=http://localhost:3080
+#==============================#
+# Static File Cache Control    #
+#==============================#
 
-###########################
-# Email
-###########################
+# Leave commented out to use defaults: 1 day (86400 seconds) for s-maxage and 2 days (172800 seconds) for max-age
+# NODE_ENV must be set to production for these to take effect
+# STATIC_CACHE_MAX_AGE=172800
+# STATIC_CACHE_S_MAX_AGE=86400
 
-# Email is used for password reset. Note that all 4 values must be set for email to work.
-# Failing to set the 4 values will result in LibreChat using the unsecured password reset!
-EMAIL_SERVICE=   # eg. gmail
-EMAIL_USERNAME=  # eg. your email address if using gmail
-EMAIL_PASSWORD=  # eg. this is the "app password" if using gmail
-EMAIL_FROM=noreply@librechat.ai # email address for from field, it is required to set a value here even in the cases where it's not porperly working.
+# If you have another service in front of your LibreChat doing compression, disable express based compression here
+# DISABLE_COMPRESSION=true
+
+#===================================================#
+#                        UI                         #
+#===================================================#
+
+APP_TITLE=LibreChat
+# CUSTOM_FOOTER="My custom footer"
+HELP_AND_FAQ_URL=https://librechat.ai
+
+# SHOW_BIRTHDAY_ICON=true
+
+# Google tag manager id
+#ANALYTICS_GTM_ID=user provided google tag manager id
+
+#===============#
+# REDIS Options #
+#===============#
+
+# REDIS_URI=10.10.10.10:6379
+# USE_REDIS=true
+
+# USE_REDIS_CLUSTER=true
+# REDIS_CA=/path/to/ca.crt
+
+#==================================================#
+#                      Others                      #
+#==================================================#
+#   You should leave the following commented out   #
+
+# NODE_ENV=
+
+# E2E_USER_EMAIL=
+# E2E_USER_PASSWORD=
+
+#=====================================================#
+#                  Cache Headers                      #
+#=====================================================#
+#   Headers that control caching of the index.html    #
+#   Default configuration prevents caching to ensure  #
+#   users always get the latest version. Customize    #
+#   only if you understand caching implications.      #
+
+# INDEX_HTML_CACHE_CONTROL=no-cache, no-store, must-revalidate
+# INDEX_HTML_PRAGMA=no-cache
+# INDEX_HTML_EXPIRES=0
+
+# no-cache: Forces validation with server before using cached version
+# no-store: Prevents storing the response entirely
+# must-revalidate: Prevents using stale content when offline
+
+#=====================================================#
+#                  OpenWeather                        #
+#=====================================================#
+OPENWEATHER_API_KEY=

.eslintrc.js

@@ -1,152 +0,0 @@
-module.exports = {
-  env: {
-    browser: true,
-    es2021: true,
-    node: true,
-    commonjs: true,
-    es6: true,
-  },
-  extends: [
-    'eslint:recommended',
-    'plugin:react/recommended',
-    'plugin:react-hooks/recommended',
-    'plugin:jest/recommended',
-    'prettier',
-  ],
-  ignorePatterns: [
-    'client/dist/**/*',
-    'client/public/**/*',
-    'e2e/playwright-report/**/*',
-    'packages/data-provider/types/**/*',
-    'packages/data-provider/dist/**/*',
-  ],
-  parser: '@typescript-eslint/parser',
-  parserOptions: {
-    ecmaVersion: 'latest',
-    sourceType: 'module',
-    ecmaFeatures: {
-      jsx: true,
-    },
-  },
-  plugins: ['react', 'react-hooks', '@typescript-eslint', 'import'],
-  rules: {
-    'react/react-in-jsx-scope': 'off',
-    '@typescript-eslint/ban-ts-comment': ['error', { 'ts-ignore': 'allow' }],
-    indent: ['error', 2, { SwitchCase: 1 }],
-    'max-len': [
-      'error',
-      {
-        code: 120,
-        ignoreStrings: true,
-        ignoreTemplateLiterals: true,
-        ignoreComments: true,
-      },
-    ],
-    'linebreak-style': 0,
-    curly: ['error', 'all'],
-    semi: ['error', 'always'],
-    'object-curly-spacing': ['error', 'always'],
-    'no-multiple-empty-lines': ['error', { max: 1 }],
-    'no-trailing-spaces': 'error',
-    'comma-dangle': ['error', 'always-multiline'],
-    // "arrow-parens": [2, "as-needed", { requireForBlockBody: true }],
-    // 'no-plusplus': ['error', { allowForLoopAfterthoughts: true }],
-    'no-console': 'off',
-    'import/no-cycle': 'error',
-    'import/no-self-import': 'error',
-    'import/extensions': 'off',
-    'no-promise-executor-return': 'off',
-    'no-param-reassign': 'off',
-    'no-continue': 'off',
-    'no-restricted-syntax': 'off',
-    'react/prop-types': ['off'],
-    'react/display-name': ['off'],
-    'no-unused-vars': ['error', { varsIgnorePattern: '^_' }],
-    quotes: ['error', 'single'],
-  },
-  overrides: [
-    {
-      files: ['**/*.ts', '**/*.tsx'],
-      rules: {
-        'no-unused-vars': 'off', // off because it conflicts with '@typescript-eslint/no-unused-vars'
-        'react/display-name': 'off',
-        '@typescript-eslint/no-unused-vars': 'warn',
-      },
-    },
-    {
-      files: ['rollup.config.js', '.eslintrc.js', 'jest.config.js'],
-      env: {
-        node: true,
-      },
-    },
-    {
-      files: [
-        '**/*.test.js',
-        '**/*.test.jsx',
-        '**/*.test.ts',
-        '**/*.test.tsx',
-        '**/*.spec.js',
-        '**/*.spec.jsx',
-        '**/*.spec.ts',
-        '**/*.spec.tsx',
-        'setupTests.js',
-      ],
-      env: {
-        jest: true,
-        node: true,
-      },
-      rules: {
-        'react/display-name': 'off',
-        'react/prop-types': 'off',
-        'react/no-unescaped-entities': 'off',
-      },
-    },
-    {
-      files: ['**/*.ts', '**/*.tsx'],
-      parser: '@typescript-eslint/parser',
-      parserOptions: {
-        project: './client/tsconfig.json',
-      },
-      plugins: ['@typescript-eslint/eslint-plugin', 'jest'],
-      extends: [
-        'plugin:@typescript-eslint/eslint-recommended',
-        'plugin:@typescript-eslint/recommended',
-      ],
-      rules: {
-        '@typescript-eslint/no-explicit-any': 'error',
-      },
-    },
-    {
-      files: './packages/data-provider/**/*.ts',
-      overrides: [
-        {
-          files: '**/*.ts',
-          parser: '@typescript-eslint/parser',
-          parserOptions: {
-            project: './packages/data-provider/tsconfig.json',
-          },
-        },
-      ],
-    },
-  ],
-  settings: {
-    react: {
-      createClass: 'createReactClass', // Regex for Component Factory to use,
-      // default to "createReactClass"
-      pragma: 'React', // Pragma to use, default to "React"
-      fragment: 'Fragment', // Fragment to use (may be a property of <pragma>), default to "Fragment"
-      version: 'detect', // React version. "detect" automatically picks the version you have installed.
-    },
-    'import/parsers': {
-      '@typescript-eslint/parser': ['.ts', '.tsx'],
-    },
-    'import/resolver': {
-      typescript: {
-        project: ['./client/tsconfig.json'],
-      },
-      node: {
-        project: ['./client/tsconfig.json'],
-      },
-    },
-  },
-};

.github/CODE_OF_CONDUCT.md

@@ -60,7 +60,7 @@ representative at an online or offline event.
 
 Instances of abusive, harassing, or otherwise unacceptable behavior may be
 reported to the community leaders responsible for enforcement here on GitHub or
-on the official [Discord Server](https://discord.gg/uDyZ5Tzhct). 
+on the official [Discord Server](https://discord.librechat.ai). 
 All complaints will be reviewed and investigated promptly and fairly.
 
 All community leaders are obligated to respect the privacy and security of the

.github/CONTRIBUTING.md

@@ -8,7 +8,7 @@ If the feature you would like to contribute has not already received prior appro
 
 Please note that a pull request involving a feature that has not been reviewed and approved by the project maintainers may be rejected. We appreciate your understanding and cooperation.
 
-If you would like to discuss the changes you wish to make, join our [Discord community](https://discord.gg/uDyZ5Tzhct), where you can engage with other contributors and seek guidance from the community.
+If you would like to discuss the changes you wish to make, join our [Discord community](https://discord.librechat.ai), where you can engage with other contributors and seek guidance from the community.
 
 ## Our Standards
 
@@ -126,6 +126,18 @@ Apply the following naming conventions to branches, labels, and other Git-relate
       
    - **Current Stance**: At present, this backend transition is of lower priority and might not be pursued.
 
+## 7. Module Import Conventions
+
+- `npm` packages first, 
+     - from shortest line (top) to longest (bottom)
+
+- Followed by typescript types (pertains to data-provider and client workspaces)
+     - longest line (top) to shortest (bottom)
+     - types from package come first
+
+- Lastly, local imports
+     - longest line (top) to shortest (bottom)
+     - imports with alias `~` treated the same as relative import with respect to line length
 
 ---
 

.github/ISSUE_TEMPLATE/BUG-REPORT.yml

@@ -1,20 +1,19 @@
 name: Bug Report
 description: File a bug report
 title: "[Bug]: "
-labels: ["bug"]
+labels: ["🐛 bug"]
 body:
   - type: markdown
     attributes:
       value: |
         Thanks for taking the time to fill out this bug report!
-  - type: input
-    id: contact
-    attributes:
-      label: Contact Details
-      description: How can we get in touch with you if we need more info?
-      placeholder: ex. email@example.com
-    validations:
-      required: false
+        
+        Before submitting, please:
+        - Search existing [Issues and Discussions](https://github.com/danny-avila/LibreChat/discussions) to see if your bug has already been reported
+        - Use [Discussions](https://github.com/danny-avila/LibreChat/discussions) instead of Issues for:
+          - General inquiries
+          - Help with setup
+          - Questions about whether you're experiencing a bug
   - type: textarea
     id: what-happened
     attributes:
@@ -23,6 +22,23 @@ body:
       placeholder: Please give as many details as possible
     validations:
       required: true
+  - type: textarea
+    id: version-info
+    attributes:
+      label: Version Information
+      description: |
+        If using Docker, please run and provide the output of:
+        ```bash
+        docker images | grep librechat
+        ```
+        
+        If running from source, please run and provide the output of:
+        ```bash
+        git rev-parse HEAD
+        ```
+      placeholder: Paste the output here
+    validations:
+      required: true
   - type: textarea
     id: steps-to-reproduce
     attributes:
@@ -47,7 +63,21 @@ body:
     id: logs
     attributes:
       label: Relevant log output
-      description: Please copy and paste any relevant log output. This will be automatically formatted into code, so no need for backticks.
+      description: |
+        Please paste relevant logs that were created when reproducing the error.
+        
+        Log locations:
+        - Docker: Project root directory ./logs
+        - npm: ./api/logs
+        
+        There are two types of logs that can help diagnose the issue:
+        - debug logs (debug-YYYY-MM-DD.log)
+        - error logs (error-YYYY-MM-DD.log)
+        
+        Error logs contain exact stack traces and are especially helpful, but both can provide valuable information.
+        Please only include the relevant portions of logs that correspond to when you reproduced the error.
+
+        For UI-related issues, browser console logs can be very helpful. You can provide these as screenshots or paste the text here.
       render: shell
   - type: textarea
     id: screenshots
@@ -58,7 +88,7 @@ body:
     id: terms
     attributes:
       label: Code of Conduct
-      description: By submitting this issue, you agree to follow our [Code of Conduct](https://github.com/danny-avila/LibreChat/blob/main/CODE_OF_CONDUCT.md)
+      description: By submitting this issue, you agree to follow our [Code of Conduct](https://github.com/danny-avila/LibreChat/blob/main/.github/CODE_OF_CONDUCT.md)
       options:
         - label: I agree to follow this project's Code of Conduct
-          required: true
+          required: true

.github/ISSUE_TEMPLATE/FEATURE-REQUEST.yml

@@ -1,20 +1,12 @@
 name: Feature Request
 description: File a feature request
-title: "Enhancement: "
-labels: ["enhancement"]
+title: "[Enhancement]: "
+labels: ["✨ enhancement"]
 body:
   - type: markdown
     attributes:
       value: |
         Thank you for taking the time to fill this out!
-  - type: input
-    id: contact
-    attributes:
-      label: Contact Details
-      description: How can we contact you if we need more information?
-      placeholder: ex. email@example.com
-    validations:
-      required: false
   - type: textarea
     id: what
     attributes:
@@ -51,7 +43,7 @@ body:
     id: terms
     attributes:
       label: Code of Conduct
-      description: By submitting this issue, you agree to follow our [Code of Conduct](https://github.com/danny-avila/LibreChat/blob/main/CODE_OF_CONDUCT.md)
+      description: By submitting this issue, you agree to follow our [Code of Conduct](https://github.com/danny-avila/LibreChat/blob/main/.github/CODE_OF_CONDUCT.md)
       options:
         - label: I agree to follow this project's Code of Conduct
           required: true

.github/ISSUE_TEMPLATE/LOCIZE_TRANSLATION_ACCESS_REQUEST.yml

@@ -0,0 +1,42 @@
+name: Locize Translation Access Request
+description: Request access to an additional language in Locize for LibreChat translations.
+title: "Locize Access Request: "
+labels: ["🌍 i18n", "🔑 access request"]
+body:
+  - type: markdown
+    attributes:
+      value: |
+        Thank you for your interest in contributing to LibreChat translations!  
+        Please fill out the form below to request access to an additional language in **Locize**.
+
+        **🔗 Available Languages:** [View the list here](https://www.librechat.ai/docs/translation)
+        
+        **📌 Note:** Ensure that the requested language is supported before submitting your request.
+  - type: input
+    id: account_name
+    attributes:
+      label: Locize Account Name
+      description: Please provide your Locize account name (e.g., John Doe).
+      placeholder: e.g., John Doe
+    validations:
+      required: true
+  - type: input
+    id: language_requested
+    attributes:
+      label: Language Code (ISO 639-1)
+      description: |
+        Enter the **ISO 639-1** language code for the language you want to translate into.  
+        Example: `es` for Spanish, `zh-Hant` for Traditional Chinese.
+        
+        **🔗 Reference:** [Available Languages](https://www.librechat.ai/docs/translation)
+      placeholder: e.g., es
+    validations:
+      required: true
+  - type: checkboxes
+    id: agreement
+    attributes:
+      label: Agreement
+      description: By submitting this request, you confirm that you will contribute responsibly and adhere to the project guidelines.
+      options:
+        - label: I agree to use my access solely for contributing to LibreChat translations.
+          required: true

.github/ISSUE_TEMPLATE/NEW-LANGUAGE-REQUEST.yml

@@ -0,0 +1,33 @@
+name: New Language Request
+description: Request to add a new language for LibreChat translations.
+title: "New Language Request: "
+labels: ["✨ enhancement", "🌍 i18n"]
+body:
+  - type: markdown
+    attributes:
+      value: |
+        Thank you for taking the time to submit a new language request! Please fill out the following details so we can review your request.
+  - type: input
+    id: language_name
+    attributes:
+      label: Language Name
+      description: Please provide the full name of the language (e.g., Spanish, Mandarin).
+      placeholder: e.g., Spanish
+    validations:
+      required: true
+  - type: input
+    id: iso_code
+    attributes:
+      label: ISO 639-1 Code
+      description: Please provide the ISO 639-1 code for the language (e.g., es for Spanish). You can refer to [this list](https://www.w3schools.com/tags/ref_language_codes.asp) for valid codes.
+      placeholder: e.g., es
+    validations:
+      required: true
+  - type: checkboxes
+    id: terms
+    attributes:
+      label: Code of Conduct
+      description: By submitting this issue, you agree to follow our [Code of Conduct](https://github.com/danny-avila/LibreChat/blob/main/.github/CODE_OF_CONDUCT.md).
+      options:
+        - label: I agree to follow this project's Code of Conduct
+          required: true

.github/ISSUE_TEMPLATE/QUESTION.yml

@@ -1,58 +0,0 @@
-name: Question
-description: Ask your question
-title: "[Question]: "
-labels: ["question"]
-body:
-  - type: markdown
-    attributes:
-      value: |
-        Thanks for taking the time to fill this!
-  - type: input
-    id: contact
-    attributes:
-      label: Contact Details
-      description: How can we get in touch with you if we need more info?
-      placeholder: ex. email@example.com
-    validations:
-      required: false
-  - type: textarea
-    id: what-is-your-question
-    attributes:
-      label: What is your question?
-      description: Please give as many details as possible
-      placeholder: Please give as many details as possible
-    validations:
-      required: true
-  - type: textarea
-    id: more-details
-    attributes:
-      label: More Details
-      description: Please provide more details if needed.
-      placeholder: Please provide more details if needed.
-    validations:
-      required: true
-  - type: dropdown
-    id: browsers
-    attributes:
-      label: What is the main subject of your question?
-      multiple: true
-      options:
-        - Documentation
-        - Installation
-        - UI
-        - Endpoints
-        - User System/OAuth
-        - Other
-  - type: textarea
-    id: screenshots
-    attributes:
-      label: Screenshots
-      description: If applicable, add screenshots to help explain your problem. You can drag and drop, paste images directly here or link to them.
-  - type: checkboxes
-    id: terms
-    attributes:
-      label: Code of Conduct
-      description: By submitting this issue, you agree to follow our [Code of Conduct](https://github.com/danny-avila/LibreChat/blob/main/CODE_OF_CONDUCT.md)
-      options:
-        - label: I agree to follow this project's Code of Conduct
-          required: true

.github/SECURITY.md

@@ -12,7 +12,7 @@ When reporting a security vulnerability, you have the following options to reach
 
 - **Option 2: GitHub Issues**: You can initiate first contact via GitHub Issues. However, please note that initial contact through GitHub Issues should not include any sensitive details.
 
-- **Option 3: Discord Server**: You can join our [Discord community](https://discord.gg/5rbRxn4uME) and initiate first contact in the `#issues` channel. However, please ensure that initial contact through Discord does not include any sensitive details.
+- **Option 3: Discord Server**: You can join our [Discord community](https://discord.librechat.ai) and initiate first contact in the `#issues` channel. However, please ensure that initial contact through Discord does not include any sensitive details.
 
 _After the initial contact, we will establish a private communication channel for further discussion._
 
@@ -39,11 +39,11 @@ Please note that as a security-conscious community, we may not always disclose d
 
 This security policy applies to the following GitHub repository:
 
-- Repository: [LibreChat](https://github.com/danny-avila/LibreChat)
+- Repository: [LibreChat](https://github.librechat.ai)
 
 ## Contact
 
-If you have any questions or concerns regarding the security of our project, please join our [Discord community](https://discord.gg/NGaa9RPCft) and report them in the appropriate channel. You can also reach out to us by [opening an issue](https://github.com/danny-avila/LibreChat/issues/new) on GitHub. Please note that the response time may vary depending on the nature and severity of the inquiry.
+If you have any questions or concerns regarding the security of our project, please join our [Discord community](https://discord.librechat.ai) and report them in the appropriate channel. You can also reach out to us by [opening an issue](https://github.com/danny-avila/LibreChat/issues/new) on GitHub. Please note that the response time may vary depending on the nature and severity of the inquiry.
 
 ## Acknowledgments
 

.github/configuration-release.json

@@ -0,0 +1,60 @@
+{
+  "categories": [
+    {
+      "title": "### ✨ New Features",
+      "labels": ["feat"]
+    },
+    {
+      "title": "### 🌍 Internationalization",
+      "labels": ["i18n"]
+    },
+    {
+      "title": "### 👐 Accessibility",
+      "labels": ["a11y"]
+    },
+    {
+      "title": "### 🔧 Fixes",
+      "labels": ["Fix", "fix"]
+    },
+    {
+      "title": "### ⚙️ Other Changes",
+      "labels": ["ci", "style", "docs", "refactor", "chore"]
+    }
+  ],
+  "ignore_labels": [
+    "🔁 duplicate",
+    "📊 analytics",
+    "🌱 good first issue",
+    "🔍 investigation",
+    "🙏 help wanted",
+    "❌ invalid",
+    "❓ question",
+    "🚫 wontfix",
+    "🚀 release",
+    "version"
+  ],
+  "base_branches": ["main"],
+  "sort": {
+    "order": "ASC",
+    "on_property": "mergedAt"
+  },
+  "label_extractor": [
+    {
+      "pattern": "^(?:[^A-Za-z0-9]*)(feat|fix|chore|docs|refactor|ci|style|a11y|i18n)\\s*:",
+      "target": "$1",
+      "flags": "i",
+      "on_property": "title",
+      "method": "match"
+    },
+    {
+      "pattern": "^(?:[^A-Za-z0-9]*)(v\\d+\\.\\d+\\.\\d+(?:-rc\\d+)?).*",
+      "target": "version",
+      "flags": "i",
+      "on_property": "title",
+      "method": "match"
+    }
+  ],
+  "template": "## [#{{TO_TAG}}] - #{{TO_TAG_DATE}}\n\nChanges from #{{FROM_TAG}} to #{{TO_TAG}}.\n\n#{{CHANGELOG}}\n\n[See full release details][release-#{{TO_TAG}}]\n\n[release-#{{TO_TAG}}]: https://github.com/#{{OWNER}}/#{{REPO}}/releases/tag/#{{TO_TAG}}\n\n---",
+  "pr_template": "- #{{TITLE}} by **@#{{AUTHOR}}** in [##{{NUMBER}}](#{{URL}})",
+  "empty_template": "- no changes"
+}

.github/configuration-unreleased.json

@@ -0,0 +1,68 @@
+{
+  "categories": [
+    {
+      "title": "### ✨ New Features",
+      "labels": ["feat"]
+    },
+    {
+      "title": "### 🌍 Internationalization",
+      "labels": ["i18n"]
+    },
+    {
+      "title": "### 👐 Accessibility",
+      "labels": ["a11y"]
+    },
+    {
+      "title": "### 🔧 Fixes",
+      "labels": ["Fix", "fix"]
+    },
+    {
+      "title": "### ⚙️ Other Changes",
+      "labels": ["ci", "style", "docs", "refactor", "chore"]
+    }
+  ],
+  "ignore_labels": [
+    "🔁 duplicate",
+    "📊 analytics",
+    "🌱 good first issue",
+    "🔍 investigation",
+    "🙏 help wanted",
+    "❌ invalid",
+    "❓ question",
+    "🚫 wontfix",
+    "🚀 release",
+    "version",
+    "action"
+  ],
+  "base_branches": ["main"],
+  "sort": {
+    "order": "ASC",
+    "on_property": "mergedAt"
+  },
+  "label_extractor": [
+    {
+      "pattern": "^(?:[^A-Za-z0-9]*)(feat|fix|chore|docs|refactor|ci|style|a11y|i18n)\\s*:",
+      "target": "$1",
+      "flags": "i",
+      "on_property": "title",
+      "method": "match"
+    },
+    {
+      "pattern": "^(?:[^A-Za-z0-9]*)(v\\d+\\.\\d+\\.\\d+(?:-rc\\d+)?).*",
+      "target": "version",
+      "flags": "i",
+      "on_property": "title",
+      "method": "match"
+    },
+    {
+      "pattern": "^(?:[^A-Za-z0-9]*)(action)\\b.*",
+      "target": "action",
+      "flags": "i",
+      "on_property": "title",
+      "method": "match"
+    }
+  ],
+  "template": "## [Unreleased]\n\n#{{CHANGELOG}}\n\n---",
+  "pr_template": "- #{{TITLE}} by **@#{{AUTHOR}}** in [##{{NUMBER}}](#{{URL}})",
+  "empty_template": "- no changes"
+}

.github/dependabot.yml

@@ -1,47 +0,0 @@
-# To get started with Dependabot version updates, you'll need to specify which
-# package ecosystems to update and where the package manifests are located.
-# Please see the documentation for all configuration options:
-# https://docs.github.com/github/administering-a-repository/configuration-options-for-dependency-updates
-
-version: 2
-updates:
-  - package-ecosystem: "npm" # See documentation for possible values
-    directory: "/api" # Location of package manifests
-    target-branch: "dev"
-    versioning-strategy: increase-if-necessary
-    schedule:
-      interval: "weekly"
-    allow:
-      # Allow both direct and indirect updates for all packages
-      - dependency-type: "all"
-    commit-message:
-      prefix: "npm api prod"
-      prefix-development: "npm api dev"
-      include: "scope"
-  - package-ecosystem: "npm" # See documentation for possible values
-    directory: "/client" # Location of package manifests
-    target-branch: "dev"
-    versioning-strategy: increase-if-necessary
-    schedule:
-      interval: "weekly"
-    allow:
-      # Allow both direct and indirect updates for all packages
-      - dependency-type: "all"
-    commit-message:
-      prefix: "npm client prod"
-      prefix-development: "npm client dev"
-      include: "scope"
-  - package-ecosystem: "npm" # See documentation for possible values
-    directory: "/" # Location of package manifests
-    target-branch: "dev"
-    versioning-strategy: increase-if-necessary
-    schedule:
-      interval: "weekly"
-    allow:
-      # Allow both direct and indirect updates for all packages
-      - dependency-type: "all"
-    commit-message:
-      prefix: "npm all prod"
-      prefix-development: "npm all dev"
-      include: "scope" 
-     

.github/playwright.yml

@@ -0,0 +1,72 @@
+# name: Playwright Tests
+# on:
+#   pull_request:
+#     branches: 
+#       - main
+#       - dev
+#       - release/*
+#     paths:
+#       - 'api/**'
+#       - 'client/**'
+#       - 'packages/**'
+#       - 'e2e/**'
+# jobs:
+#   tests_e2e:
+#     name: Run Playwright tests
+#     if: github.event.pull_request.head.repo.full_name == 'danny-avila/LibreChat'
+#     timeout-minutes: 60
+#     runs-on: ubuntu-latest
+#     env:
+#       NODE_ENV: CI
+#       CI: true
+#       SEARCH: false
+#       BINGAI_TOKEN: user_provided
+#       CHATGPT_TOKEN: user_provided
+#       MONGO_URI: ${{ secrets.MONGO_URI }}
+#       OPENAI_API_KEY: ${{ secrets.OPENAI_API_KEY }}
+#       E2E_USER_EMAIL: ${{ secrets.E2E_USER_EMAIL }}
+#       E2E_USER_PASSWORD: ${{ secrets.E2E_USER_PASSWORD }}
+#       JWT_SECRET: ${{ secrets.JWT_SECRET }}
+#       JWT_REFRESH_SECRET: ${{ secrets.JWT_REFRESH_SECRET }}
+#       CREDS_KEY: ${{ secrets.CREDS_KEY }}
+#       CREDS_IV: ${{ secrets.CREDS_IV }}
+#       DOMAIN_CLIENT: ${{ secrets.DOMAIN_CLIENT }}
+#       DOMAIN_SERVER: ${{ secrets.DOMAIN_SERVER }}
+#       PLAYWRIGHT_SKIP_BROWSER_DOWNLOAD: 1 # Skip downloading during npm install
+#       PLAYWRIGHT_BROWSERS_PATH: 0 # Places binaries to node_modules/@playwright/test
+#       TITLE_CONVO: false
+#     steps:
+#       - uses: actions/checkout@v4
+#       - uses: actions/setup-node@v4
+#         with:
+#           node-version: 18
+#           cache: 'npm'
+
+#       - name: Install global dependencies
+#         run: npm ci
+
+#       # - name: Remove sharp dependency
+#       #   run: rm -rf node_modules/sharp
+
+#       # - name: Install sharp with linux dependencies
+#       #   run: cd api && SHARP_IGNORE_GLOBAL_LIBVIPS=1 npm install --arch=x64 --platform=linux --libc=glibc sharp
+
+#       - name: Build Client
+#         run: npm run frontend
+
+#       - name: Install Playwright
+#         run: |
+#           npx playwright install-deps
+#           npm install -D @playwright/test@latest
+#           npx playwright install chromium
+
+#       - name: Run Playwright tests
+#         run: npm run e2e:ci
+
+#       - name: Upload playwright report
+#         uses: actions/upload-artifact@v3
+#         if: always()
+#         with:
+#           name: playwright-report
+#           path: e2e/playwright-report/
+#           retention-days: 30

.github/pull_request_template.md

@@ -1,7 +1,10 @@
 # Pull Request Template
 
+⚠️ Before Submitting a PR, Please Review:
+- Please ensure that you have thoroughly read and understood the [Contributing Docs](https://github.com/danny-avila/LibreChat/blob/main/.github/CONTRIBUTING.md) before submitting your Pull Request.
 
-### ⚠️ Before Submitting a PR, read the [Contributing Docs](./CONTRIBUTING.md) in full!
+⚠️ Documentation Updates Notice:
+- Kindly note that documentation updates are managed in this repository: [librechat.ai](https://github.com/LibreChat-AI/librechat.ai)
 
 ## Summary
 
@@ -15,7 +18,7 @@ Please delete any irrelevant options.
 - [ ] New feature (non-breaking change which adds functionality)
 - [ ] Breaking change (fix or feature that would cause existing functionality to not work as expected)
 - [ ] This change requires a documentation update
-- [ ] Documentation update 
+- [ ] Translation update
 
 ## Testing
 
@@ -25,6 +28,8 @@ Please describe your test process and include instructions so that we can reprod
 
 ## Checklist
 
+Please delete any irrelevant options.
+
 - [ ] My code adheres to this project's style guidelines
 - [ ] I have performed a self-review of my own code
 - [ ] I have commented in any complex areas of my code
@@ -33,3 +38,4 @@ Please describe your test process and include instructions so that we can reprod
 - [ ] I have written tests demonstrating that my changes are effective or that my feature works
 - [ ] Local unit tests pass with my changes
 - [ ] Any changes dependent on mine have been merged and published in downstream modules.
+- [ ] A pull request for updating the documentation has been submitted.

.github/workflows/a11y.yml

@@ -0,0 +1,26 @@
+name: Lint for accessibility issues
+
+on:
+  pull_request:
+    paths:
+      - 'client/src/**'
+  workflow_dispatch:
+    inputs:
+      run_workflow:
+        description: 'Set to true to run this workflow'
+        required: true
+        default: 'false'
+
+jobs:
+  axe-linter:
+    runs-on: ubuntu-latest
+    if: >
+      (github.event_name == 'pull_request' && github.event.pull_request.head.repo.full_name == 'danny-avila/LibreChat') ||
+      (github.event_name == 'workflow_dispatch' && github.event.inputs.run_workflow == 'true')
+    
+    steps:
+      - uses: actions/checkout@v4
+      - uses: dequelabs/axe-linter-action@v1
+        with:
+          api_key: ${{ secrets.AXE_LINTER_API_KEY }}
+          github_token: ${{ secrets.GITHUB_TOKEN }}

.github/workflows/backend-review.yml

@@ -23,9 +23,9 @@ jobs:
       BAN_INTERVAL: ${{ secrets.BAN_INTERVAL }}
       NODE_ENV: CI
     steps:
-      - uses: actions/checkout@v2
+      - uses: actions/checkout@v4
       - name: Use Node.js 20.x
-        uses: actions/setup-node@v3
+        uses: actions/setup-node@v4
         with:
           node-version: 20
           cache: 'npm'
@@ -33,13 +33,32 @@ jobs:
       - name: Install dependencies
         run: npm ci
 
-      - name: Install Data Provider
+      - name: Install Data Provider Package
         run: npm run build:data-provider
 
+      - name: Install MCP Package
+        run: npm run build:mcp
+
+      - name: Create empty auth.json file
+        run: |
+          mkdir -p api/data
+          echo '{}' > api/data/auth.json
+
+      - name: Check for Circular dependency in rollup
+        working-directory: ./packages/data-provider
+        run: |
+          output=$(npm run rollup:api)
+          echo "$output"
+          if echo "$output" | grep -q "Circular dependency"; then
+            echo "Error: Circular dependency detected!"
+            exit 1
+          fi
+
+      - name: Prepare .env.test file
+        run: cp api/test/.env.test.example api/test/.env.test
+
       - name: Run unit tests
         run: cd api && npm run test:ci
 
-      - name: Run linters
-        uses: wearerequired/lint-action@v2
-        with:
-          eslint: true
+      - name: Run librechat-data-provider unit tests
+        run: cd packages/data-provider && npm run test:ci

.github/workflows/container.yml

@@ -1,52 +0,0 @@
-name: Docker Compose Build on Tag
-
-# The workflow is triggered when a tag is pushed
-on:
-  push:
-    tags:
-      - "*"
-
-jobs:
-  build:
-    runs-on: ubuntu-latest
-
-    steps:
-      # Check out the repository
-      - name: Checkout
-        uses: actions/checkout@v2
-
-      # Set up Docker
-      - name: Set up Docker
-        uses: docker/setup-buildx-action@v1
-
-      # Log in to GitHub Container Registry
-      - name: Log in to GitHub Container Registry
-        uses: docker/login-action@v2
-        with:
-          registry: ghcr.io
-          username: ${{ github.actor }}
-          password: ${{ secrets.GITHUB_TOKEN }}
-
-      # Run docker-compose build
-      - name: Build Docker images
-        run: |
-          cp .env.example .env
-          docker-compose build
-          docker build -f Dockerfile.multi --target api-build -t librechat-api .
-
-      # Get Tag Name
-      - name: Get Tag Name
-        id: tag_name
-        run: echo "TAG_NAME=${GITHUB_REF/refs\/tags\//}" >> $GITHUB_ENV
-
-      # Tag it properly before push to github
-      - name: tag image and push
-        run: |
-          docker tag librechat:latest ghcr.io/${{ github.repository_owner }}/librechat:${{ env.TAG_NAME }}
-          docker push ghcr.io/${{ github.repository_owner }}/librechat:${{ env.TAG_NAME }}
-          docker tag librechat:latest ghcr.io/${{ github.repository_owner }}/librechat:latest
-          docker push ghcr.io/${{ github.repository_owner }}/librechat:latest
-          docker tag librechat-api:latest ghcr.io/${{ github.repository_owner }}/librechat-api:${{ env.TAG_NAME }}
-          docker push ghcr.io/${{ github.repository_owner }}/librechat-api:${{ env.TAG_NAME }}
-          docker tag librechat-api:latest ghcr.io/${{ github.repository_owner }}/librechat-api:latest
-          docker push ghcr.io/${{ github.repository_owner }}/librechat-api:latest

.github/workflows/data-provider.yml

@@ -11,8 +11,8 @@ jobs:
   build:
     runs-on: ubuntu-latest
     steps:
-      - uses: actions/checkout@v3
-      - uses: actions/setup-node@v3
+      - uses: actions/checkout@v4
+      - uses: actions/setup-node@v4
         with:
           node-version: 16
       - run: cd packages/data-provider && npm ci
@@ -22,8 +22,8 @@ jobs:
     needs: build
     runs-on: ubuntu-latest
     steps:
-      - uses: actions/checkout@v3
-      - uses: actions/setup-node@v3
+      - uses: actions/checkout@v4
+      - uses: actions/setup-node@v4
         with:
           node-version: 16
           registry-url: 'https://registry.npmjs.org'

.github/workflows/deploy-dev.yml

@@ -0,0 +1,41 @@
+name: Update Test Server
+
+on:
+  workflow_run:
+    workflows: ["Docker Dev Images Build"]
+    types:
+      - completed
+  workflow_dispatch:
+
+jobs:
+  deploy:
+    runs-on: ubuntu-latest
+    if: |
+      github.repository == 'danny-avila/LibreChat' &&
+      (github.event_name == 'workflow_dispatch' || github.event.workflow_run.conclusion == 'success')
+    steps:
+    - name: Checkout repository
+      uses: actions/checkout@v4
+
+    - name: Install SSH Key
+      uses: shimataro/ssh-key-action@v2
+      with:
+        key: ${{ secrets.DO_SSH_PRIVATE_KEY }}
+        known_hosts: ${{ secrets.DO_KNOWN_HOSTS }}
+
+    - name: Run update script on DigitalOcean Droplet
+      env:
+        DO_HOST: ${{ secrets.DO_HOST }}
+        DO_USER: ${{ secrets.DO_USER }}
+      run: |
+        ssh -o StrictHostKeyChecking=no ${DO_USER}@${DO_HOST} << EOF
+        sudo -i -u danny bash << EEOF
+        cd ~/LibreChat && \
+        git fetch origin main && \
+        npm run update:deployed && \
+        git checkout do-deploy && \
+        git rebase main && \
+        npm run start:deployed && \
+        echo "Update completed. Application should be running now."
+        EEOF
+        EOF

.github/workflows/deploy.yml

@@ -17,7 +17,7 @@ jobs:
     steps:
       # checkout the repo
       - name: 'Checkout GitHub Action'
-        uses: actions/checkout@main
+        uses: actions/checkout@v4
 
       - name: 'Login via Azure CLI'
         uses: azure/login@v1

.github/workflows/dev-images.yml

@@ -8,19 +8,33 @@ on:
     paths:
       - 'api/**'
       - 'client/**'
+      - 'packages/**'
 
 jobs:
   build:
     runs-on: ubuntu-latest
+    strategy:
+      matrix:
+        include:
+          - target: api-build
+            file: Dockerfile.multi
+            image_name: librechat-dev-api
+          - target: node
+            file: Dockerfile
+            image_name: librechat-dev
 
     steps:
       # Check out the repository
       - name: Checkout
-        uses: actions/checkout@v2
+        uses: actions/checkout@v4
 
-      # Set up Docker
-      - name: Set up Docker
-        uses: docker/setup-buildx-action@v1
+      # Set up QEMU
+      - name: Set up QEMU
+        uses: docker/setup-qemu-action@v3
+
+      # Set up Docker Buildx
+      - name: Set up Docker Buildx
+        uses: docker/setup-buildx-action@v3
 
       # Log in to GitHub Container Registry
       - name: Log in to GitHub Container Registry
@@ -30,22 +44,29 @@ jobs:
           username: ${{ github.actor }}
           password: ${{ secrets.GITHUB_TOKEN }}
 
-      # Build Docker images
-      - name: Build Docker images
+      # Login to Docker Hub
+      - name: Login to Docker Hub
+        uses: docker/login-action@v3
+        with:
+          username: ${{ secrets.DOCKERHUB_USERNAME }}
+          password: ${{ secrets.DOCKERHUB_TOKEN }}
+
+      # Prepare the environment
+      - name: Prepare environment
         run: |
           cp .env.example .env
-          docker build -f Dockerfile.multi --target api-build -t librechat-dev-api .
-          docker build -f Dockerfile -t librechat-dev .
 
-      # Tag and push the images to GitHub Container Registry
-      - name: Tag and push images
-        run: |
-          docker tag librechat-dev-api:latest ghcr.io/${{ github.repository_owner }}/librechat-dev-api:${{ github.sha }}
-          docker push ghcr.io/${{ github.repository_owner }}/librechat-dev-api:${{ github.sha }}
-          docker tag librechat-dev-api:latest ghcr.io/${{ github.repository_owner }}/librechat-dev-api:latest
-          docker push ghcr.io/${{ github.repository_owner }}/librechat-dev-api:latest
-
-          docker tag librechat-dev:latest ghcr.io/${{ github.repository_owner }}/librechat-dev:${{ github.sha }}
-          docker push ghcr.io/${{ github.repository_owner }}/librechat-dev:${{ github.sha }}
-          docker tag librechat-dev:latest ghcr.io/${{ github.repository_owner }}/librechat-dev:latest
-          docker push ghcr.io/${{ github.repository_owner }}/librechat-dev:latest
+      # Build and push Docker images for each target
+      - name: Build and push Docker images
+        uses: docker/build-push-action@v5
+        with:
+          context: .
+          file: ${{ matrix.file }}
+          push: true
+          tags: |
+            ghcr.io/${{ github.repository_owner }}/${{ matrix.image_name }}:${{ github.sha }}
+            ghcr.io/${{ github.repository_owner }}/${{ matrix.image_name }}:latest
+            ${{ secrets.DOCKERHUB_USERNAME }}/${{ matrix.image_name }}:${{ github.sha }}
+            ${{ secrets.DOCKERHUB_USERNAME }}/${{ matrix.image_name }}:latest
+          platforms: linux/amd64,linux/arm64
+          target: ${{ matrix.target }}

.github/workflows/eslint-ci.yml

@@ -0,0 +1,73 @@
+name: ESLint Code Quality Checks
+
+on:
+  pull_request:
+    branches:
+      - main
+      - dev
+      - release/*
+    paths:
+      - 'api/**'
+      - 'client/**'
+
+jobs:
+  eslint_checks:
+    name: Run ESLint Linting
+    runs-on: ubuntu-latest
+    permissions:
+      contents: read
+      security-events: write
+      actions: read
+    steps:
+      - name: Checkout repository
+        uses: actions/checkout@v4
+        with:
+          fetch-depth: 0
+
+      - name: Set up Node.js 20.x
+        uses: actions/setup-node@v4
+        with:
+          node-version: 20
+          cache: npm
+
+      - name: Install dependencies
+        run: npm ci
+
+      # Run ESLint on changed files within the api/ and client/ directories.
+      - name: Run ESLint on changed files
+        env:
+          SARIF_ESLINT_IGNORE_SUPPRESSED: "true"
+        run: |
+          # Extract the base commit SHA from the pull_request event payload.
+          BASE_SHA=$(jq --raw-output .pull_request.base.sha "$GITHUB_EVENT_PATH")
+          echo "Base commit SHA: $BASE_SHA"
+
+          # Get changed files (only JS/TS files in api/ or client/)
+          CHANGED_FILES=$(git diff --name-only --diff-filter=ACMRTUXB "$BASE_SHA" HEAD | grep -E '^(api|client)/.*\.(js|jsx|ts|tsx)$' || true)
+
+          # Debug output
+          echo "Changed files:"
+          echo "$CHANGED_FILES"
+
+          # Ensure there are files to lint before running ESLint
+          if [[ -z "$CHANGED_FILES" ]]; then
+            echo "No matching files changed. Skipping ESLint."
+            echo "UPLOAD_SARIF=false" >> $GITHUB_ENV
+            exit 0
+          fi
+
+          # Set variable to allow SARIF upload
+          echo "UPLOAD_SARIF=true" >> $GITHUB_ENV
+
+          # Run ESLint
+          npx eslint --no-error-on-unmatched-pattern \
+            --config eslint.config.mjs \
+            --format @microsoft/eslint-formatter-sarif \
+            --output-file eslint-results.sarif $CHANGED_FILES || true
+
+      - name: Upload analysis results to GitHub
+        if: env.UPLOAD_SARIF == 'true'
+        uses: github/codeql-action/upload-sarif@v3
+        with:
+          sarif_file: eslint-results.sarif
+          wait-for-processing: true

.github/workflows/frontend-review.yml

@@ -1,11 +1,6 @@
-#github action to run unit tests for frontend with jest
 name: Frontend Unit Tests
+
 on:
-  # push:
-  #   branches: 
-  #     - main
-  #     - dev
-  #     - release/*
   pull_request:
     branches: 
       - main
@@ -14,15 +9,38 @@ on:
     paths:
       - 'client/**'
       - 'packages/**'
+
 jobs:
-  tests_frontend:
-    name: Run frontend unit tests
+  tests_frontend_ubuntu:
+    name: Run frontend unit tests on Ubuntu
     timeout-minutes: 60
     runs-on: ubuntu-latest
     steps:
-      - uses: actions/checkout@v2
+      - uses: actions/checkout@v4
       - name: Use Node.js 20.x
-        uses: actions/setup-node@v3
+        uses: actions/setup-node@v4
+        with:
+          node-version: 20
+          cache: 'npm'
+
+      - name: Install dependencies
+        run: npm ci
+
+      - name: Build Client
+        run: npm run frontend:ci
+
+      - name: Run unit tests
+        run: npm run test:ci --verbose
+        working-directory: client
+
+  tests_frontend_windows:
+    name: Run frontend unit tests on Windows
+    timeout-minutes: 60
+    runs-on: windows-latest
+    steps:
+      - uses: actions/checkout@v4
+      - name: Use Node.js 20.x
+        uses: actions/setup-node@v4
         with:
           node-version: 20
           cache: 'npm'

.github/workflows/generate-release-changelog-pr.yml

@@ -0,0 +1,94 @@
+name: Generate Release Changelog PR
+
+on:
+  push:
+    tags:
+      - 'v*.*.*'
+
+jobs:
+  generate-release-changelog-pr:
+    permissions:
+      contents: write    # Needed for pushing commits and creating branches.
+      pull-requests: write
+    runs-on: ubuntu-latest
+    steps:
+      # 1. Checkout the repository (with full history).
+      - name: Checkout Repository
+        uses: actions/checkout@v4
+        with:
+          fetch-depth: 0
+
+      # 2. Generate the release changelog using our custom configuration.
+      - name: Generate Release Changelog
+        id: generate_release
+        uses: mikepenz/release-changelog-builder-action@v5.1.0
+        env:
+          GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
+        with:
+          configuration: ".github/configuration-release.json"
+          owner: ${{ github.repository_owner }}
+          repo: ${{ github.event.repository.name }}
+          outputFile: CHANGELOG-release.md
+
+      # 3. Update the main CHANGELOG.md:
+      #    - If it doesn't exist, create it with a basic header.
+      #    - Remove the "Unreleased" section (if present).
+      #    - Prepend the new release changelog above previous releases.
+      #    - Remove all temporary files before committing.
+      - name: Update CHANGELOG.md
+        run: |
+          # Determine the release tag, e.g. "v1.2.3"
+          TAG=${GITHUB_REF##*/}
+          echo "Using release tag: $TAG"
+          
+          # Ensure CHANGELOG.md exists; if not, create a basic header.
+          if [ ! -f CHANGELOG.md ]; then
+            echo "# Changelog" > CHANGELOG.md
+            echo "" >> CHANGELOG.md
+            echo "All notable changes to this project will be documented in this file." >> CHANGELOG.md
+            echo "" >> CHANGELOG.md
+          fi
+
+          echo "Updating CHANGELOG.md…"
+          
+          # Remove the "Unreleased" section (from "## [Unreleased]" until the first occurrence of '---') if it exists.
+          if grep -q "^## \[Unreleased\]" CHANGELOG.md; then
+            awk '/^## \[Unreleased\]/{flag=1} flag && /^---/{flag=0; next} !flag' CHANGELOG.md > CHANGELOG.cleaned
+          else
+            cp CHANGELOG.md CHANGELOG.cleaned
+          fi
+
+          # Split the cleaned file into:
+          #   - header.md: content before the first release header ("## [v...").
+          #   - tail.md: content from the first release header onward.
+          awk '/^## \[v/{exit} {print}' CHANGELOG.cleaned > header.md
+          awk 'f{print} /^## \[v/{f=1; print}' CHANGELOG.cleaned > tail.md
+
+          # Combine header, the new release changelog, and the tail.
+          echo "Combining updated changelog parts..."
+          cat header.md CHANGELOG-release.md > CHANGELOG.md.new
+          echo "" >> CHANGELOG.md.new
+          cat tail.md >> CHANGELOG.md.new
+
+          mv CHANGELOG.md.new CHANGELOG.md
+
+          # Remove temporary files.
+          rm -f CHANGELOG.cleaned header.md tail.md CHANGELOG-release.md
+
+          echo "Final CHANGELOG.md content:"
+          cat CHANGELOG.md
+
+      # 4. Create (or update) the Pull Request with the updated CHANGELOG.md.
+      - name: Create Pull Request
+        uses: peter-evans/create-pull-request@v7
+        with:
+          token: ${{ secrets.GITHUB_TOKEN }}
+          sign-commits: true
+          commit-message: "chore: update CHANGELOG for release ${GITHUB_REF##*/}"
+          base: main
+          branch: "changelog/${GITHUB_REF##*/}"
+          reviewers: danny-avila
+          title: "chore: update CHANGELOG for release ${GITHUB_REF##*/}"
+          body: |
+            **Description**:
+            - This PR updates the CHANGELOG.md by removing the "Unreleased" section and adding new release notes for release ${GITHUB_REF##*/} above previous releases.

.github/workflows/generate-unreleased-changelog-pr.yml

@@ -0,0 +1,106 @@
+name: Generate Unreleased Changelog PR
+
+on:
+  schedule:
+    - cron: "0 0 * * 1"  # Runs every Monday at 00:00 UTC
+
+jobs:
+  generate-unreleased-changelog-pr:
+    permissions:
+      contents: write    # Needed for pushing commits and creating branches.
+      pull-requests: write
+    runs-on: ubuntu-latest
+    steps:
+      # 1. Checkout the repository on main.
+      - name: Checkout Repository on Main
+        uses: actions/checkout@v4
+        with:
+          ref: main
+          fetch-depth: 0
+
+      # 4. Get the latest version tag.
+      - name: Get Latest Tag
+        id: get_latest_tag
+        run: |
+          LATEST_TAG=$(git describe --tags $(git rev-list --tags --max-count=1) || echo "none")
+          echo "Latest tag: $LATEST_TAG"
+          echo "tag=$LATEST_TAG" >> $GITHUB_OUTPUT
+
+      # 5. Generate the Unreleased changelog.
+      - name: Generate Unreleased Changelog
+        id: generate_unreleased
+        uses: mikepenz/release-changelog-builder-action@v5.1.0
+        env:
+          GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
+        with:
+          configuration: ".github/configuration-unreleased.json"
+          owner: ${{ github.repository_owner }}
+          repo: ${{ github.event.repository.name }}
+          outputFile: CHANGELOG-unreleased.md
+          fromTag: ${{ steps.get_latest_tag.outputs.tag }}
+          toTag: main
+
+      # 7. Update CHANGELOG.md with the new Unreleased section.
+      - name: Update CHANGELOG.md
+        id: update_changelog
+        run: |
+          # Create CHANGELOG.md if it doesn't exist.
+          if [ ! -f CHANGELOG.md ]; then
+            echo "# Changelog" > CHANGELOG.md
+            echo "" >> CHANGELOG.md
+            echo "All notable changes to this project will be documented in this file." >> CHANGELOG.md
+            echo "" >> CHANGELOG.md
+          fi
+          
+          echo "Updating CHANGELOG.md…"
+          
+          # Extract content before the "## [Unreleased]" (or first version header if missing).
+          if grep -q "^## \[Unreleased\]" CHANGELOG.md; then
+            awk '/^## \[Unreleased\]/{exit} {print}' CHANGELOG.md > CHANGELOG_TMP.md
+          else
+            awk '/^## \[v/{exit} {print}' CHANGELOG.md > CHANGELOG_TMP.md
+          fi
+          
+          # Append the generated Unreleased changelog.
+          echo "" >> CHANGELOG_TMP.md
+          cat CHANGELOG-unreleased.md >> CHANGELOG_TMP.md
+          echo "" >> CHANGELOG_TMP.md
+          
+          # Append the remainder of the original changelog (starting from the first version header).
+          awk 'f{print} /^## \[v/{f=1; print}' CHANGELOG.md >> CHANGELOG_TMP.md
+          
+          # Replace the old file with the updated file.
+          mv CHANGELOG_TMP.md CHANGELOG.md
+          
+          # Remove the temporary generated file.
+          rm -f CHANGELOG-unreleased.md
+          
+          echo "Final CHANGELOG.md:"
+          cat CHANGELOG.md
+
+      # 8. Check if CHANGELOG.md has any updates.
+      - name: Check for CHANGELOG.md changes
+        id: changelog_changes
+        run: |
+          if git diff --quiet CHANGELOG.md; then
+            echo "has_changes=false" >> $GITHUB_OUTPUT
+          else
+            echo "has_changes=true" >> $GITHUB_OUTPUT
+          fi
+
+      # 9. Create (or update) the Pull Request only if there are changes.
+      - name: Create Pull Request
+        if: steps.changelog_changes.outputs.has_changes == 'true'
+        uses: peter-evans/create-pull-request@v7
+        with:
+          token: ${{ secrets.GITHUB_TOKEN }}
+          base: main
+          branch: "changelog/unreleased-update"
+          sign-commits: true
+          commit-message: "action: update Unreleased changelog"
+          title: "action: update Unreleased changelog"
+          body: |
+            **Description**:
+            - This PR updates the Unreleased section in CHANGELOG.md.
+            - It compares the current main branch with the latest version tag (determined as ${{ steps.get_latest_tag.outputs.tag }}),
+              regenerates the Unreleased changelog, removes any old Unreleased block, and inserts the new content.

.github/workflows/generate_embeddings.yml

@@ -0,0 +1,20 @@
+name: 'generate_embeddings'
+on:
+  workflow_dispatch:
+  push:
+    branches:
+      - main
+    paths:
+      - 'docs/**'
+
+jobs:
+  generate:
+    runs-on: ubuntu-latest
+    steps:
+      - uses: actions/checkout@v3
+      - uses: supabase/embeddings-generator@v0.0.5
+        with:
+          supabase-url: ${{ secrets.SUPABASE_URL }}
+          supabase-service-role-key: ${{ secrets.SUPABASE_SERVICE_ROLE_KEY }}
+          openai-key: ${{ secrets.OPENAI_DOC_EMBEDDINGS_KEY }}
+          docs-root-path: 'docs'

.github/workflows/helmcharts.yml

@@ -0,0 +1,33 @@
+name: Build Helm Charts on Tag
+
+# The workflow is triggered when a tag is pushed
+on:
+  push:
+    tags:
+      - "*"
+
+jobs:
+  release:
+    permissions:
+      contents: write
+    runs-on: ubuntu-latest
+    steps:
+      - name: Checkout
+        uses: actions/checkout@v4
+        with:
+          fetch-depth: 0
+
+      - name: Configure Git
+        run: |
+          git config user.name "$GITHUB_ACTOR"
+          git config user.email "$GITHUB_ACTOR@users.noreply.github.com"
+
+      - name: Install Helm
+        uses: azure/setup-helm@v4
+        env:
+          GITHUB_TOKEN: "${{ secrets.GITHUB_TOKEN }}"
+
+      - name: Run chart-releaser
+        uses: helm/chart-releaser-action@v1.6.0
+        env:
+          CR_TOKEN: "${{ secrets.GITHUB_TOKEN }}"

.github/workflows/i18n-unused-keys.yml

@@ -0,0 +1,93 @@
+name: Detect Unused i18next Strings
+
+on:
+  pull_request:
+    paths:
+      - "client/src/**"
+      - "api/**"
+
+jobs:
+  detect-unused-i18n-keys:
+    runs-on: ubuntu-latest
+    permissions:
+      pull-requests: write  # Required for posting PR comments
+    steps:
+      - name: Checkout repository
+        uses: actions/checkout@v3
+
+      - name: Find unused i18next keys
+        id: find-unused
+        run: |
+          echo "🔍 Scanning for unused i18next keys..."
+
+          # Define paths
+          I18N_FILE="client/src/locales/en/translation.json"
+          SOURCE_DIRS=("client/src" "api")
+
+          # Check if translation file exists
+          if [[ ! -f "$I18N_FILE" ]]; then
+            echo "::error title=Missing i18n File::Translation file not found: $I18N_FILE"
+            exit 1
+          fi
+
+          # Extract all keys from the JSON file
+          KEYS=$(jq -r 'keys[]' "$I18N_FILE")
+
+          # Track unused keys
+          UNUSED_KEYS=()
+
+          # Check if each key is used in the source code
+          for KEY in $KEYS; do
+            FOUND=false
+            for DIR in "${SOURCE_DIRS[@]}"; do
+              if grep -r --include=\*.{js,jsx,ts,tsx} -q "$KEY" "$DIR"; then
+                FOUND=true
+                break
+              fi
+            done
+          
+            if [[ "$FOUND" == false ]]; then
+              UNUSED_KEYS+=("$KEY")
+            fi
+          done
+
+          # Output results
+          if [[ ${#UNUSED_KEYS[@]} -gt 0 ]]; then
+            echo "🛑 Found ${#UNUSED_KEYS[@]} unused i18n keys:"
+            echo "unused_keys=$(echo "${UNUSED_KEYS[@]}" | jq -R -s -c 'split(" ")')" >> $GITHUB_ENV
+            for KEY in "${UNUSED_KEYS[@]}"; do
+              echo "::warning title=Unused i18n Key::'$KEY' is defined but not used in the codebase."
+            done
+          else
+            echo "✅ No unused i18n keys detected!"
+            echo "unused_keys=[]" >> $GITHUB_ENV
+          fi
+
+      - name: Post verified comment on PR
+        if: env.unused_keys != '[]'
+        run: |
+          PR_NUMBER=$(jq --raw-output .pull_request.number "$GITHUB_EVENT_PATH")
+
+          # Format the unused keys list as checkboxes for easy manual checking.
+          FILTERED_KEYS=$(echo "$unused_keys" | jq -r '.[]' | grep -v '^\s*$' | sed 's/^/- [ ] `/;s/$/`/' )
+
+          COMMENT_BODY=$(cat <<EOF
+          ### 🚨 Unused i18next Keys Detected
+
+          The following translation keys are defined in \`translation.json\` but are **not used** in the codebase:
+
+          $FILTERED_KEYS
+
+          ⚠️ **Please remove these unused keys to keep the translation files clean.**
+          EOF
+          )
+
+          gh api "repos/${{ github.repository }}/issues/${PR_NUMBER}/comments" \
+            -f body="$COMMENT_BODY" \
+            -H "Authorization: token ${{ secrets.GITHUB_TOKEN }}"
+        env:
+          GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
+
+      - name: Fail workflow if unused keys found
+        if: env.unused_keys != '[]'
+        run: exit 1

.github/workflows/latest-images-main.yml

@@ -1,40 +0,0 @@
-name: Docker Compose Build on Main Branch
-
-on:
-  workflow_dispatch:  # This line allows manual triggering
-
-jobs:
-  build:
-    runs-on: ubuntu-latest
-
-    steps:
-      # Check out the repository
-      - name: Checkout
-        uses: actions/checkout@v2
-
-      # Set up Docker
-      - name: Set up Docker
-        uses: docker/setup-buildx-action@v1
-
-      # Log in to GitHub Container Registry
-      - name: Log in to GitHub Container Registry
-        uses: docker/login-action@v2
-        with:
-          registry: ghcr.io
-          username: ${{ github.actor }}
-          password: ${{ secrets.GITHUB_TOKEN }}
-
-      # Run docker-compose build
-      - name: Build Docker images
-        run: |
-          cp .env.example .env
-          docker-compose build
-          docker build -f Dockerfile.multi --target api-build -t librechat-api .
-
-      # Tag and push the images with the 'latest' tag
-      - name: Tag image and push
-        run: |
-          docker tag librechat:latest ghcr.io/${{ github.repository_owner }}/librechat:latest
-          docker push ghcr.io/${{ github.repository_owner }}/librechat:latest
-          docker tag librechat-api:latest ghcr.io/${{ github.repository_owner }}/librechat-api:latest
-          docker push ghcr.io/${{ github.repository_owner }}/librechat-api:latest

.github/workflows/locize-i18n-sync.yml

@@ -0,0 +1,72 @@
+name: Sync Locize Translations & Create Translation PR
+
+on:
+  push:
+    branches: [main]
+  repository_dispatch:
+    types: [locize/versionPublished]
+
+jobs:
+  sync-translations:
+    name: Sync Translation Keys with Locize
+    runs-on: ubuntu-latest
+    steps:
+      - name: Checkout Repository
+        uses: actions/checkout@v4
+
+      - name: Set Up Node.js
+        uses: actions/setup-node@v4
+        with:
+          node-version: 20
+
+      - name: Install locize CLI
+        run: npm install -g locize-cli
+
+      # Sync translations (Push missing keys & remove deleted ones)
+      - name: Sync Locize with Repository
+        if: ${{ github.event_name == 'push' }}
+        run: |
+          cd client/src/locales
+          locize sync --api-key ${{ secrets.LOCIZE_API_KEY }} --project-id ${{ secrets.LOCIZE_PROJECT_ID }} --language en
+
+      # When triggered by repository_dispatch, skip sync step.
+      - name: Skip sync step on non-push events
+        if: ${{ github.event_name != 'push' }}
+        run: echo "Skipping sync as the event is not a push."
+
+  create-pull-request:
+    name: Create Translation PR on Version Published
+    runs-on: ubuntu-latest
+    needs: sync-translations
+    permissions:
+      contents: write
+      pull-requests: write
+    steps:
+      # 1. Check out the repository.
+      - name: Checkout Repository
+        uses: actions/checkout@v4
+
+      # 2. Download translation files from locize.
+      - name: Download Translations from locize
+        uses: locize/download@v1
+        with:
+          project-id: ${{ secrets.LOCIZE_PROJECT_ID }}
+          path: "client/src/locales"
+
+      # 3. Create a Pull Request using built-in functionality.
+      - name: Create Pull Request
+        uses: peter-evans/create-pull-request@v7
+        with:
+          token: ${{ secrets.GITHUB_TOKEN }}
+          sign-commits: true
+          commit-message: "🌍 i18n: Update translation.json with latest translations"
+          base: main
+          branch: i18n/locize-translation-update
+          reviewers: danny-avila
+          title: "🌍 i18n: Update translation.json with latest translations"
+          body: |
+            **Description**:
+            - 🎯 **Objective**: Update `translation.json` with the latest translations from locize.
+            - 🔍 **Details**: This PR is automatically generated upon receiving a versionPublished event with version "latest". It reflects the newest translations provided by locize.
+            - ✅ **Status**: Ready for review.
+          labels: "🌍 i18n"

.github/workflows/main-image-workflow.yml

@@ -0,0 +1,69 @@
+name: Docker Compose Build Latest Main Image Tag (Manual Dispatch)
+
+on:
+  workflow_dispatch:
+
+jobs:
+  build:
+    runs-on: ubuntu-latest
+    strategy:
+      matrix:
+        include:
+          - target: api-build
+            file: Dockerfile.multi
+            image_name: librechat-api
+          - target: node
+            file: Dockerfile
+            image_name: librechat
+
+    steps:
+      - name: Checkout
+        uses: actions/checkout@v4
+
+      - name: Fetch tags and set the latest tag
+        run: |
+          git fetch --tags
+          echo "LATEST_TAG=$(git describe --tags `git rev-list --tags --max-count=1`)" >> $GITHUB_ENV
+
+      # Set up QEMU
+      - name: Set up QEMU
+        uses: docker/setup-qemu-action@v3
+
+      # Set up Docker Buildx
+      - name: Set up Docker Buildx
+        uses: docker/setup-buildx-action@v3
+
+      # Log in to GitHub Container Registry
+      - name: Log in to GitHub Container Registry
+        uses: docker/login-action@v2
+        with:
+          registry: ghcr.io
+          username: ${{ github.actor }}
+          password: ${{ secrets.GITHUB_TOKEN }}
+
+      # Login to Docker Hub
+      - name: Login to Docker Hub
+        uses: docker/login-action@v3
+        with:
+          username: ${{ secrets.DOCKERHUB_USERNAME }}
+          password: ${{ secrets.DOCKERHUB_TOKEN }}
+
+      # Prepare the environment
+      - name: Prepare environment
+        run: |
+          cp .env.example .env
+
+      # Build and push Docker images for each target
+      - name: Build and push Docker images
+        uses: docker/build-push-action@v5
+        with:
+          context: .
+          file: ${{ matrix.file }}
+          push: true
+          tags: |
+            ghcr.io/${{ github.repository_owner }}/${{ matrix.image_name }}:${{ env.LATEST_TAG }}
+            ghcr.io/${{ github.repository_owner }}/${{ matrix.image_name }}:latest
+            ${{ secrets.DOCKERHUB_USERNAME }}/${{ matrix.image_name }}:${{ env.LATEST_TAG }}
+            ${{ secrets.DOCKERHUB_USERNAME }}/${{ matrix.image_name }}:latest
+          platforms: linux/amd64,linux/arm64
+          target: ${{ matrix.target }}

.github/workflows/mkdocs.yaml

@@ -1,24 +0,0 @@
-name: mkdocs
-on:
-  push:
-    branches:
-      - main
-permissions:
-  contents: write
-jobs:
-  deploy:
-    runs-on: ubuntu-latest
-    steps:
-      - uses: actions/checkout@v3
-      - uses: actions/setup-python@v4
-        with:
-          python-version: 3.x
-      - run: echo "cache_id=$(date --utc '+%V')" >> $GITHUB_ENV 
-      - uses: actions/cache@v3
-        with:
-          key: mkdocs-material-${{ env.cache_id }}
-          path: .cache
-          restore-keys: |
-            mkdocs-material-
-      - run: pip install mkdocs-material 
-      - run: mkdocs gh-deploy --force

.github/workflows/playwright.yml

@@ -1,72 +0,0 @@
-name: Playwright Tests
-on:
-  pull_request:
-    branches: 
-      - main
-      - dev
-      - release/*
-    paths:
-      - 'api/**'
-      - 'client/**'
-      - 'packages/**'
-      - 'e2e/**'
-jobs:
-  tests_e2e:
-    name: Run Playwright tests
-    if: github.event.pull_request.head.repo.full_name == 'danny-avila/LibreChat'
-    timeout-minutes: 60
-    runs-on: ubuntu-latest
-    env:
-      NODE_ENV: CI
-      CI: true
-      SEARCH: false
-      BINGAI_TOKEN: user_provided
-      CHATGPT_TOKEN: user_provided
-      MONGO_URI: ${{ secrets.MONGO_URI }}
-      OPENAI_API_KEY: ${{ secrets.OPENAI_API_KEY }}
-      E2E_USER_EMAIL: ${{ secrets.E2E_USER_EMAIL }}
-      E2E_USER_PASSWORD: ${{ secrets.E2E_USER_PASSWORD }}
-      JWT_SECRET: ${{ secrets.JWT_SECRET }}
-      JWT_REFRESH_SECRET: ${{ secrets.JWT_REFRESH_SECRET }}
-      CREDS_KEY: ${{ secrets.CREDS_KEY }}
-      CREDS_IV: ${{ secrets.CREDS_IV }}
-      DOMAIN_CLIENT: ${{ secrets.DOMAIN_CLIENT }}
-      DOMAIN_SERVER: ${{ secrets.DOMAIN_SERVER }}
-      PLAYWRIGHT_SKIP_BROWSER_DOWNLOAD: 1 # Skip downloading during npm install
-      PLAYWRIGHT_BROWSERS_PATH: 0 # Places binaries to node_modules/@playwright/test
-      TITLE_CONVO: false
-    steps:
-      - uses: actions/checkout@v3
-      - uses: actions/setup-node@v3
-        with:
-          node-version: 18
-          cache: 'npm'
-
-      - name: Install global dependencies
-        run: npm ci
-
-      # - name: Remove sharp dependency
-      #   run: rm -rf node_modules/sharp
-
-      # - name: Install sharp with linux dependencies
-      #   run: cd api && SHARP_IGNORE_GLOBAL_LIBVIPS=1 npm install --arch=x64 --platform=linux --libc=glibc sharp
-
-      - name: Build Client
-        run: npm run frontend
-
-      - name: Install Playwright
-        run: |
-          npx playwright install-deps
-          npm install -D @playwright/test@latest
-          npx playwright install chromium
-
-      - name: Run Playwright tests
-        run: npm run e2e:ci
-
-      - name: Upload playwright report
-        uses: actions/upload-artifact@v3
-        if: always()
-        with:
-          name: playwright-report
-          path: e2e/playwright-report/
-          retention-days: 30

.github/workflows/tag-images.yml

@@ -0,0 +1,67 @@
+name: Docker Images Build on Tag
+
+on:
+  push:
+    tags:
+      - '*'
+
+jobs:
+  build:
+    runs-on: ubuntu-latest
+    strategy:
+      matrix:
+        include:
+          - target: api-build
+            file: Dockerfile.multi
+            image_name: librechat-api
+          - target: node
+            file: Dockerfile
+            image_name: librechat
+
+    steps:
+      # Check out the repository
+      - name: Checkout
+        uses: actions/checkout@v4
+
+      # Set up QEMU
+      - name: Set up QEMU
+        uses: docker/setup-qemu-action@v3
+
+      # Set up Docker Buildx
+      - name: Set up Docker Buildx
+        uses: docker/setup-buildx-action@v3
+
+      # Log in to GitHub Container Registry
+      - name: Log in to GitHub Container Registry
+        uses: docker/login-action@v2
+        with:
+          registry: ghcr.io
+          username: ${{ github.actor }}
+          password: ${{ secrets.GITHUB_TOKEN }}
+
+      # Login to Docker Hub
+      - name: Login to Docker Hub
+        uses: docker/login-action@v3
+        with:
+          username: ${{ secrets.DOCKERHUB_USERNAME }}
+          password: ${{ secrets.DOCKERHUB_TOKEN }}
+
+      # Prepare the environment
+      - name: Prepare environment
+        run: |
+          cp .env.example .env
+
+      # Build and push Docker images for each target
+      - name: Build and push Docker images
+        uses: docker/build-push-action@v5
+        with:
+          context: .
+          file: ${{ matrix.file }}
+          push: true
+          tags: |
+            ghcr.io/${{ github.repository_owner }}/${{ matrix.image_name }}:${{ github.ref_name }}
+            ghcr.io/${{ github.repository_owner }}/${{ matrix.image_name }}:latest
+            ${{ secrets.DOCKERHUB_USERNAME }}/${{ matrix.image_name }}:${{ github.ref_name }}
+            ${{ secrets.DOCKERHUB_USERNAME }}/${{ matrix.image_name }}:latest
+          platforms: linux/amd64,linux/arm64
+          target: ${{ matrix.target }}

.github/workflows/unused-packages.yml

@@ -0,0 +1,153 @@
+name: Detect Unused NPM Packages
+
+on:
+  pull_request:
+    paths:
+      - 'package.json'
+      - 'package-lock.json'
+      - 'client/**'
+      - 'api/**'
+
+jobs:
+  detect-unused-packages:
+    runs-on: ubuntu-latest
+    permissions:
+      pull-requests: write
+
+    steps:
+      - uses: actions/checkout@v4
+
+      - name: Use Node.js 20.x
+        uses: actions/setup-node@v4
+        with:
+          node-version: 20
+          cache: 'npm'
+
+      - name: Install depcheck
+        run: npm install -g depcheck
+
+      - name: Validate JSON files
+        run: |
+          for FILE in package.json client/package.json api/package.json; do
+            if [[ -f "$FILE" ]]; then
+              jq empty "$FILE" || (echo "::error title=Invalid JSON::$FILE is invalid" && exit 1)
+            fi
+          done
+
+      - name: Extract Dependencies Used in Scripts
+        id: extract-used-scripts
+        run: |
+          extract_deps_from_scripts() {
+            local package_file=$1
+            if [[ -f "$package_file" ]]; then
+              jq -r '.scripts | to_entries[].value' "$package_file" | \
+                grep -oE '([a-zA-Z0-9_-]+)' | sort -u > used_scripts.txt
+            else
+              touch used_scripts.txt
+            fi
+          }
+
+          extract_deps_from_scripts "package.json"
+          mv used_scripts.txt root_used_deps.txt
+
+          extract_deps_from_scripts "client/package.json"
+          mv used_scripts.txt client_used_deps.txt
+
+          extract_deps_from_scripts "api/package.json"
+          mv used_scripts.txt api_used_deps.txt
+
+      - name: Extract Dependencies Used in Source Code
+        id: extract-used-code
+        run: |
+          extract_deps_from_code() {
+            local folder=$1
+            local output_file=$2
+            if [[ -d "$folder" ]]; then
+              grep -rEho "require\\(['\"]([a-zA-Z0-9@/._-]+)['\"]\\)" "$folder" --include=\*.{js,ts,mjs,cjs} | \
+                sed -E "s/require\\(['\"]([a-zA-Z0-9@/._-]+)['\"]\\)/\1/" > "$output_file"
+          
+              grep -rEho "import .* from ['\"]([a-zA-Z0-9@/._-]+)['\"]" "$folder" --include=\*.{js,ts,mjs,cjs} | \
+                sed -E "s/import .* from ['\"]([a-zA-Z0-9@/._-]+)['\"]/\1/" >> "$output_file"
+          
+              sort -u "$output_file" -o "$output_file"
+            else
+              touch "$output_file"
+            fi
+          }
+
+          extract_deps_from_code "." root_used_code.txt
+          extract_deps_from_code "client" client_used_code.txt
+          extract_deps_from_code "api" api_used_code.txt
+
+      - name: Run depcheck for root package.json
+        id: check-root
+        run: |
+          if [[ -f "package.json" ]]; then
+            UNUSED=$(depcheck --json | jq -r '.dependencies | join("\n")' || echo "")
+            UNUSED=$(comm -23 <(echo "$UNUSED" | sort) <(cat root_used_deps.txt root_used_code.txt | sort) || echo "")
+            echo "ROOT_UNUSED<<EOF" >> $GITHUB_ENV
+            echo "$UNUSED" >> $GITHUB_ENV
+            echo "EOF" >> $GITHUB_ENV
+          fi
+
+      - name: Run depcheck for client/package.json
+        id: check-client
+        run: |
+          if [[ -f "client/package.json" ]]; then
+            chmod -R 755 client
+            cd client
+            UNUSED=$(depcheck --json | jq -r '.dependencies | join("\n")' || echo "")
+            UNUSED=$(comm -23 <(echo "$UNUSED" | sort) <(cat ../client_used_deps.txt ../client_used_code.txt | sort) || echo "")
+            echo "CLIENT_UNUSED<<EOF" >> $GITHUB_ENV
+            echo "$UNUSED" >> $GITHUB_ENV
+            echo "EOF" >> $GITHUB_ENV
+            cd ..
+          fi
+
+      - name: Run depcheck for api/package.json
+        id: check-api
+        run: |
+          if [[ -f "api/package.json" ]]; then
+            chmod -R 755 api
+            cd api
+            UNUSED=$(depcheck --json | jq -r '.dependencies | join("\n")' || echo "")
+            UNUSED=$(comm -23 <(echo "$UNUSED" | sort) <(cat ../api_used_deps.txt ../api_used_code.txt | sort) || echo "")
+            echo "API_UNUSED<<EOF" >> $GITHUB_ENV
+            echo "$UNUSED" >> $GITHUB_ENV
+            echo "EOF" >> $GITHUB_ENV
+            cd ..
+          fi
+
+      - name: Post comment on PR if unused dependencies are found
+        if: env.ROOT_UNUSED != '' || env.CLIENT_UNUSED != '' || env.API_UNUSED != ''
+        run: |
+          PR_NUMBER=$(jq --raw-output .pull_request.number "$GITHUB_EVENT_PATH")
+
+          ROOT_LIST=$(echo "$ROOT_UNUSED" | awk '{print "- `" $0 "`"}')
+          CLIENT_LIST=$(echo "$CLIENT_UNUSED" | awk '{print "- `" $0 "`"}')
+          API_LIST=$(echo "$API_UNUSED" | awk '{print "- `" $0 "`"}')
+
+          COMMENT_BODY=$(cat <<EOF
+          ### 🚨 Unused NPM Packages Detected
+
+          The following **unused dependencies** were found:
+
+          $(if [[ ! -z "$ROOT_UNUSED" ]]; then echo "#### 📂 Root \`package.json\`"; echo ""; echo "$ROOT_LIST"; echo ""; fi)
+
+          $(if [[ ! -z "$CLIENT_UNUSED" ]]; then echo "#### 📂 Client \`client/package.json\`"; echo ""; echo "$CLIENT_LIST"; echo ""; fi)
+
+          $(if [[ ! -z "$API_UNUSED" ]]; then echo "#### 📂 API \`api/package.json\`"; echo ""; echo "$API_LIST"; echo ""; fi)
+
+          ⚠️ **Please remove these unused dependencies to keep your project clean.**
+          EOF
+          )
+
+          gh api "repos/${{ github.repository }}/issues/${PR_NUMBER}/comments" \
+            -f body="$COMMENT_BODY" \
+            -H "Authorization: token ${{ secrets.GITHUB_TOKEN }}"
+        env:
+          GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
+
+      - name: Fail workflow if unused dependencies found
+        if: env.ROOT_UNUSED != '' || env.CLIENT_UNUSED != '' || env.API_UNUSED != ''
+        run: exit 1

.gitignore

@@ -2,7 +2,7 @@
 
 # Logs
 data-node
-meili_data
+meili_data*
 data/
 logs
 *.log
@@ -11,6 +11,7 @@ logs
 pids
 *.pid
 *.seed
+.git
 
 # Directory for instrumented libs generated by jscoverage/JSCover
 lib-cov
@@ -21,6 +22,10 @@ coverage
 # Grunt intermediate storage (http://gruntjs.com/creating-plugins#storing-task-files)
 .grunt
 
+# translation services
+config/translations/stores/*
+client/src/localization/languages/*_missing_keys.json
+
 # Compiled Dirs (http://nodejs.org/api/addons.html)
 build/
 dist/
@@ -40,7 +45,8 @@ meili_data/
 api/node_modules/
 client/node_modules/
 bower_components/
-types/
+*.d.ts
+!vite-env.d.ts
 
 # Floobits
 .floo
@@ -48,6 +54,10 @@ types/
 .floo
 .flooignore
 
+#config file
+librechat.yaml
+librechat.yml
+
 # Environment
 .npmrc
 .env*
@@ -65,11 +75,20 @@ src/style - official.css
 /playwright/.cache/
 .DS_Store
 *.code-workspace
+.idx
+monospace.json
 .idea
+*.iml
 *.pem
 config.local.ts
 **/storageState.json
 junit.xml
+**/.venv/
+**/venv/
+
+# docker override file
+docker-compose.override.yaml
+docker-compose.override.yml
 
 # meilisearch
 meilisearch
@@ -78,4 +97,13 @@ data.ms/*
 auth.json
 
 /packages/ux-shared/
-/images
+/images
+
+!client/src/components/Nav/SettingsTabs/Data/
+
+# User uploads
+uploads/
+
+# owner
+release/
+!/client/src/@types/i18next.d.ts

.husky/pre-commit

@@ -1,4 +1,4 @@
-#!/usr/bin/env sh
+#!/usr/bin/env sh 
 set -e
 . "$(dirname -- "$0")/_/husky.sh"
 [ -n "$CI" ] && exit 0

.prettierrc

@@ -0,0 +1,19 @@
+{
+  "tailwindConfig": "./client/tailwind.config.mjs",
+  "printWidth": 100,
+  "tabWidth": 2,
+  "useTabs": false,
+  "semi": true,
+  "singleQuote": true,
+  "trailingComma": "all",
+  "arrowParens": "always",
+  "embeddedLanguageFormatting": "auto",
+  "insertPragma": false,
+  "proseWrap": "preserve",
+  "quoteProps": "as-needed",
+  "requirePragma": false,
+  "rangeStart": 0,
+  "endOfLine": "auto",
+  "jsxSingleQuote": false,
+  "plugins": ["prettier-plugin-tailwindcss"]
+}

.vscode/launch.json

@@ -0,0 +1,17 @@
+{
+  "version": "0.2.0",
+  "configurations": [
+    {
+      "type": "node",
+      "request": "launch",
+      "name": "Launch LibreChat (debug)",
+      "skipFiles": ["<node_internals>/**"],
+      "program": "${workspaceFolder}/api/server/index.js",
+      "env": {
+        "NODE_ENV": "production"
+      },
+      "console": "integratedTerminal",
+      "envFile": "${workspaceFolder}/.env"
+    }
+  ]
+}

Dockerfile

@@ -1,21 +1,32 @@
-# Base node image
-FROM node:19-alpine AS node
+# v0.7.7-rc1
 
-COPY . /app
+# Base node image
+FROM node:20-alpine AS node
+
+RUN apk --no-cache add curl
+
+RUN mkdir -p /app && chown node:node /app
 WORKDIR /app
 
-# Install call deps - Install curl for health check
-RUN apk --no-cache add curl && \
-    # We want to inherit env from the container, not the file
-    # This will preserve any existing env file if it's already in souce
-    # otherwise it will create a new one
-    touch .env && \
-    # Build deps in seperate 
-    npm ci
+USER node
 
-# React client build
-ENV NODE_OPTIONS="--max-old-space-size=2048"
-RUN npm run frontend
+COPY --chown=node:node . .
+
+RUN \
+    # Allow mounting of these files, which have no default
+    touch .env ; \
+    # Create directories for the volumes to inherit the correct permissions
+    mkdir -p /app/client/public/images /app/api/logs ; \
+    npm config set fetch-retry-maxtimeout 600000 ; \
+    npm config set fetch-retries 5 ; \
+    npm config set fetch-retry-mintimeout 15000 ; \
+    npm install --no-audit; \
+    # React client build
+    NODE_OPTIONS="--max-old-space-size=2048" npm run frontend; \
+    npm prune --production; \
+    npm cache clean --force
+
+RUN mkdir -p /app/client/public/images /app/api/logs
 
 # Node API setup
 EXPOSE 3080

Dockerfile.multi

@@ -1,40 +1,56 @@
-# Build API, Client and Data Provider
-FROM node:19-alpine AS base
+# Dockerfile.multi
+# v0.7.7-rc1
 
+# Base for all builds
+FROM node:20-alpine AS base-min
 WORKDIR /app
-COPY config/loader.js ./config/
-RUN npm install dotenv
+RUN apk --no-cache add curl
+RUN npm config set fetch-retry-maxtimeout 600000 && \
+    npm config set fetch-retries 5 && \
+    npm config set fetch-retry-mintimeout 15000
+COPY package*.json ./
+COPY packages/data-provider/package*.json ./packages/data-provider/
+COPY packages/mcp/package*.json ./packages/mcp/
+COPY client/package*.json ./client/
+COPY api/package*.json ./api/
 
-WORKDIR /app/api
-COPY api/package*.json ./
-COPY api/ ./
-RUN npm install
+# Install all dependencies for every build
+FROM base-min AS base
+WORKDIR /app
+RUN npm ci
 
-# React client build
+# Build data-provider
+FROM base AS data-provider-build
+WORKDIR /app/packages/data-provider
+COPY packages/data-provider ./
+RUN npm run build
+
+# Build mcp package
+FROM base AS mcp-build
+WORKDIR /app/packages/mcp
+COPY packages/mcp ./
+COPY --from=data-provider-build /app/packages/data-provider/dist /app/packages/data-provider/dist
+RUN npm run build
+
+# Client build
 FROM base AS client-build
 WORKDIR /app/client
-COPY ./client/ ./
-
-WORKDIR /app/packages/data-provider
-COPY ./packages/data-provider ./
-RUN npm install
-RUN npm run build
-RUN mkdir -p /app/client/node_modules/librechat-data-provider/
-RUN cp -R /app/packages/data-provider/* /app/client/node_modules/librechat-data-provider/
-
-WORKDIR /app/client
-RUN npm install
+COPY client ./
+COPY --from=data-provider-build /app/packages/data-provider/dist /app/packages/data-provider/dist
 ENV NODE_OPTIONS="--max-old-space-size=2048"
 RUN npm run build
 
-# Node API setup
-FROM base AS api-build
-COPY --from=client-build /app/client/dist /app/client/dist
+# API setup (including client dist)
+FROM base-min AS api-build
+WORKDIR /app
+# Install only production deps
+RUN npm ci --omit=dev
+COPY api ./api
+COPY config ./config
+COPY --from=data-provider-build /app/packages/data-provider/dist ./packages/data-provider/dist
+COPY --from=mcp-build /app/packages/mcp/dist ./packages/mcp/dist
+COPY --from=client-build /app/client/dist ./client/dist
+WORKDIR /app/api
 EXPOSE 3080
 ENV HOST=0.0.0.0
 CMD ["node", "server/index.js"]
-
-# Nginx setup
-FROM nginx:1.21.1-alpine AS prod-stage
-COPY ./client/nginx.conf /etc/nginx/conf.d/default.conf
-CMD ["nginx", "-g", "daemon off;"]

LICENSE

@@ -1,6 +1,6 @@
 MIT License
 
-Copyright (c) 2023 LibreChat
+Copyright (c) 2025 LibreChat
 
 Permission is hereby granted, free of charge, to any person obtaining a copy
 of this software and associated documentation files (the "Software"), to deal

README.md

@@ -1,14 +1,14 @@
 <p align="center">
-  <a href="https://docs.librechat.ai">
-    <img src="docs/assets/LibreChat.svg" height="256">
-  </a>
-  <a href="https://docs.librechat.ai">
-    <h1 align="center">LibreChat</h1>
+  <a href="https://librechat.ai">
+    <img src="client/public/assets/logo.svg" height="256">
   </a>
+  <h1 align="center">
+    <a href="https://librechat.ai">LibreChat</a>
+  </h1>
 </p>
 
 <p align="center">
-  <a href="https://discord.gg/NGaa9RPCft"> 
+  <a href="https://discord.librechat.ai"> 
     <img
       src="https://img.shields.io/discord/1086345563026489514?label=&logo=discord&style=for-the-badge&logoWidth=20&logoColor=white&labelColor=000000&color=blueviolet">
   </a>
@@ -20,145 +20,183 @@
     <img
       src="https://img.shields.io/badge/DOCS-blue.svg?style=for-the-badge&logo=read-the-docs&logoColor=white&labelColor=000000&logoWidth=20">
   </a>
-  <a aria-label="Sponsors" href="#sponsors">
+  <a aria-label="Sponsors" href="https://github.com/sponsors/danny-avila">
     <img
       src="https://img.shields.io/badge/SPONSORS-brightgreen.svg?style=for-the-badge&logo=github-sponsors&logoColor=white&labelColor=000000&logoWidth=20">
   </a>
 </p>
 
-## All-In-One AI Conversations with LibreChat ##
+<p align="center">
+<a href="https://railway.app/template/b5k2mn?referralCode=HI9hWz">
+  <img src="https://railway.app/button.svg" alt="Deploy on Railway" height="30">
+</a>
+<a href="https://zeabur.com/templates/0X2ZY8">
+  <img src="https://zeabur.com/button.svg" alt="Deploy on Zeabur" height="30"/>
+</a>
+<a href="https://template.cloud.sealos.io/deploy?templateName=librechat">
+  <img src="https://raw.githubusercontent.com/labring-actions/templates/main/Deploy-on-Sealos.svg" alt="Deploy on Sealos" height="30">
+</a>
+</p>
+
+<p align="center">
+  <a href="https://www.librechat.ai/docs/translation">
+    <img 
+      src="https://img.shields.io/badge/dynamic/json.svg?style=for-the-badge&color=2096F3&label=locize&query=%24.translatedPercentage&url=https://api.locize.app/badgedata/4cb2598b-ed4d-469c-9b04-2ed531a8cb45&suffix=%+translated" 
+      alt="Translation Progress">
+  </a>
+</p>
+
+
+# ✨ Features
+
+- 🖥️ **UI & Experience** inspired by ChatGPT with enhanced design and features
+
+- 🤖 **AI Model Selection**:  
+  - Anthropic (Claude), AWS Bedrock, OpenAI, Azure OpenAI, Google, Vertex AI, OpenAI Assistants API (incl. Azure)
+  - [Custom Endpoints](https://www.librechat.ai/docs/quick_start/custom_endpoints): Use any OpenAI-compatible API with LibreChat, no proxy required
+  - Compatible with [Local & Remote AI Providers](https://www.librechat.ai/docs/configuration/librechat_yaml/ai_endpoints):
+    - Ollama, groq, Cohere, Mistral AI, Apple MLX, koboldcpp, together.ai,
+    - OpenRouter, Perplexity, ShuttleAI, Deepseek, Qwen, and more
+
+- 🔧 **[Code Interpreter API](https://www.librechat.ai/docs/features/code_interpreter)**: 
+  - Secure, Sandboxed Execution in Python, Node.js (JS/TS), Go, C/C++, Java, PHP, Rust, and Fortran
+  - Seamless File Handling: Upload, process, and download files directly
+  - No Privacy Concerns: Fully isolated and secure execution
+
+- 🔦 **Agents & Tools Integration**:  
+  - **[LibreChat Agents](https://www.librechat.ai/docs/features/agents)**:
+    - No-Code Custom Assistants: Build specialized, AI-driven helpers without coding  
+    - Flexible & Extensible: Attach tools like DALL-E-3, file search, code execution, and more  
+    - Compatible with Custom Endpoints, OpenAI, Azure, Anthropic, AWS Bedrock, and more
+    - [Model Context Protocol (MCP) Support](https://modelcontextprotocol.io/clients#librechat) for Tools
+  - Use LibreChat Agents and OpenAI Assistants with Files, Code Interpreter, Tools, and API Actions
+
+- 🪄 **Generative UI with Code Artifacts**:  
+  - [Code Artifacts](https://youtu.be/GfTj7O4gmd0?si=WJbdnemZpJzBrJo3) allow creation of React, HTML, and Mermaid diagrams directly in chat
+
+- 💾 **Presets & Context Management**:  
+  - Create, Save, & Share Custom Presets  
+  - Switch between AI Endpoints and Presets mid-chat
+  - Edit, Resubmit, and Continue Messages with Conversation branching  
+  - [Fork Messages & Conversations](https://www.librechat.ai/docs/features/fork) for Advanced Context control
+
+- 💬 **Multimodal & File Interactions**:  
+  - Upload and analyze images with Claude 3, GPT-4o, o1, Llama-Vision, and Gemini 📸  
+  - Chat with Files using Custom Endpoints, OpenAI, Azure, Anthropic, AWS Bedrock, & Google 🗃️
+
+- 🌎 **Multilingual UI**:  
+  - English, 中文, Deutsch, Español, Français, Italiano, Polski, Português Brasileiro
+  - Русский, 日本語, Svenska, 한국어, Tiếng Việt, 繁體中文, العربية, Türkçe, Nederlands, עברית
+
+- 🧠 **Reasoning UI**:  
+  - Dynamic Reasoning UI for Chain-of-Thought/Reasoning AI models like DeepSeek-R1
+
+- 🎨 **Customizable Interface**:  
+  - Customizable Dropdown & Interface that adapts to both power users and newcomers
+
+- 🗣️ **Speech & Audio**:  
+  - Chat hands-free with Speech-to-Text and Text-to-Speech  
+  - Automatically send and play Audio  
+  - Supports OpenAI, Azure OpenAI, and Elevenlabs
+
+- 📥 **Import & Export Conversations**:  
+  - Import Conversations from LibreChat, ChatGPT, Chatbot UI  
+  - Export conversations as screenshots, markdown, text, json
+
+- 🔍 **Search & Discovery**:  
+  - Search all messages/conversations
+
+- 👥 **Multi-User & Secure Access**:
+  - Multi-User, Secure Authentication with OAuth2, LDAP, & Email Login Support
+  - Built-in Moderation, and Token spend tools
+
+- ⚙️ **Configuration & Deployment**:  
+  - Configure Proxy, Reverse Proxy, Docker, & many Deployment options  
+  - Use completely local or deploy on the cloud
+
+- 📖 **Open-Source & Community**:  
+  - Completely Open-Source & Built in Public  
+  - Community-driven development, support, and feedback
+
+[For a thorough review of our features, see our docs here](https://docs.librechat.ai/) 📚
+
+## 🪶 All-In-One AI Conversations with LibreChat
+
 LibreChat brings together the future of assistant AIs with the revolutionary technology of OpenAI's ChatGPT. Celebrating the original styling, LibreChat gives you the ability to integrate multiple AI models. It also integrates and enhances original client features such as conversation and message search, prompt templates and plugins.
 
 With LibreChat, you no longer need to opt for ChatGPT Plus and can instead use free or pay-per-call APIs. We welcome contributions, cloning, and forking to enhance the capabilities of this advanced chatbot platform.
-  
-<!-- https://github.com/danny-avila/LibreChat/assets/110412045/c1eb0c0f-41f6-4335-b982-84b278b53d59 -->
 
-[![Watch the video](https://img.youtube.com/vi/pNIOs1ovsXw/maxresdefault.jpg)](https://youtu.be/pNIOs1ovsXw)
+[![Watch the video](https://raw.githubusercontent.com/LibreChat-AI/librechat.ai/main/public/images/changelog/v0.7.6.gif)](https://www.youtube.com/watch?v=ilfwGQtJNlI)
+
 Click on the thumbnail to open the video☝️
 
-# Features
-- Response streaming identical to ChatGPT through server-sent events
-- UI from original ChatGPT, including Dark mode
-- AI model selection: OpenAI API, BingAI, ChatGPT Browser, PaLM2, Anthropic (Claude), Plugins
-- Create, Save, & Share custom presets - [More info on prompt presets here](https://github.com/danny-avila/LibreChat/releases/tag/v0.3.0)
-- Edit and Resubmit messages with conversation branching
-- Search all messages/conversations - [More info here](https://github.com/danny-avila/LibreChat/releases/tag/v0.1.0)
-- Plugins now available (including web access, image generation and more)
+---
+
+## 🌐 Resources
+
+**GitHub Repo:**
+  - **RAG API:** [github.com/danny-avila/rag_api](https://github.com/danny-avila/rag_api)
+  - **Website:** [github.com/LibreChat-AI/librechat.ai](https://github.com/LibreChat-AI/librechat.ai)
+
+**Other:**
+  - **Website:** [librechat.ai](https://librechat.ai)
+  - **Documentation:** [docs.librechat.ai](https://docs.librechat.ai)
+  - **Blog:** [blog.librechat.ai](https://blog.librechat.ai)
 
 ---
 
-## ⚠️ [Breaking Changes](docs/general_info/breaking_changes.md) ⚠️
+## 📝 Changelog
 
-**Please read this before updating from a previous version**
+Keep up with the latest updates by visiting the releases page and notes:
+- [Releases](https://github.com/danny-avila/LibreChat/releases)
+- [Changelog](https://www.librechat.ai/changelog) 
+
+**⚠️ Please consult the [changelog](https://www.librechat.ai/changelog) for breaking changes before updating.**
 
 ---
 
-## Changelog 
-Keep up with the latest updates by visiting the releases page - [Releases](https://github.com/danny-avila/LibreChat/releases)
+## ⭐ Star History
+
+<p align="center">
+  <a href="https://star-history.com/#danny-avila/LibreChat&Date">
+    <img alt="Star History Chart" src="https://api.star-history.com/svg?repos=danny-avila/LibreChat&type=Date&theme=dark" onerror="this.src='https://api.star-history.com/svg?repos=danny-avila/LibreChat&type=Date'" />
+  </a>
+</p>
+<p align="center">
+  <a href="https://trendshift.io/repositories/4685" target="_blank" style="padding: 10px;">
+    <img src="https://trendshift.io/api/badge/repositories/4685" alt="danny-avila%2FLibreChat | Trendshift" style="width: 250px; height: 55px;" width="250" height="55"/>
+  </a>
+  <a href="https://runacap.com/ross-index/q1-24/" target="_blank" rel="noopener" style="margin-left: 20px;">
+    <img style="width: 260px; height: 56px" src="https://runacap.com/wp-content/uploads/2024/04/ROSS_badge_white_Q1_2024.svg" alt="ROSS Index - Fastest Growing Open-Source Startups in Q1 2024 | Runa Capital" width="260" height="56"/>
+  </a>
+</p>
 
 ---
 
-<h1>Table of Contents</h1>
+## ✨ Contributions
 
-<details open>
-  <summary><strong>Getting Started</strong></summary>
+Contributions, suggestions, bug reports and fixes are welcome!
 
-  * Installation
-    * [Docker Compose Install🐳](docs/install/docker_compose_install.md)
-    * [Linux Install🐧](docs/install/linux_install.md)
-    * [Mac Install🍎](docs/install/mac_install.md)
-    * [Windows Install💙](docs/install/windows_install.md)
-  * Configuration
-    * [APIs and Tokens](docs/install/apis_and_tokens.md)
-    * [User Auth System](docs/install/user_auth_system.md)
-    * [Online MongoDB Database](docs/install/mongodb.md)
-    * [Default Language](docs/install/default_language.md)
-</details>
-
-<details>
-  <summary><strong>General Information</strong></summary>
-
-  * [Code of Conduct](.github/CODE_OF_CONDUCT.md)
-  * [Project Origin](docs/general_info/project_origin.md)
-  * [Multilingual Information](docs/general_info/multilingual_information.md)
-  * [Tech Stack](docs/general_info/tech_stack.md)   
-</details>
-
-<details>
-  <summary><strong>Features</strong></summary>
-
-  * **Plugins**
-    * [Introduction](docs/features/plugins/introduction.md)
-    * [Google](docs/features/plugins/google_search.md)
-    * [Stable Diffusion](docs/features/plugins/stable_diffusion.md)
-    * [Wolfram](docs/features/plugins/wolfram.md)
-    * [Make Your Own Plugin](docs/features/plugins/make_your_own.md)
-    * [Using official ChatGPT Plugins](docs/features/plugins/chatgpt_plugins_openapi.md)
-
-  
-  * [Automated Moderation](docs/features/mod_system.md)
-  * [Third-Party Tools](docs/features/third_party.md)
-  * [Proxy](docs/features/proxy.md)
-  * [Bing Jailbreak](docs/features/bing_jailbreak.md)
-  * [Token Usage](docs/features/token_usage.md)
-</details>
-
-<details>
-  <summary><strong>Cloud Deployment</strong></summary>
-
-  * [DigitalOcean](docs/deployment/digitalocean.md)
-  * [Azure](docs/deployment/azure-terraform.md)
-  * [Linode](docs/deployment/linode.md)
-  * [Cloudflare](docs/deployment/cloudflare.md)
-  * [Ngrok](docs/deployment/ngrok.md)
-  * [HuggingFace](docs/deployment/huggingface.md)
-  * [Render](docs/deployment/render.md)
-  * [Meilisearch in Render](docs/deployment/meilisearch_in_render.md)
-  * [Hetzner](docs/deployment/hetzner_ubuntu.md)
-  * [Heroku](docs/deployment/heroku.md)
-</details>
-
-<details>
-  <summary><strong>Contributions</strong></summary>
-  
-  * [Contributor Guidelines](.github/CONTRIBUTING.md)
-  * [Documentation Guidelines](docs/contributions/documentation_guidelines.md)
-  * [Contribute a Translation](docs/contributions/translation_contribution.md)
-  * [Code Standards and Conventions](docs/contributions/coding_conventions.md)
-  * [Testing](docs/contributions/testing.md)
-  * [Security](.github/SECURITY.md)
-  * [Project Roadmap](https://github.com/users/danny-avila/projects/2)
-</details>
+For new features, components, or extensions, please open an issue and discuss before sending a PR.
 
+If you'd like to help translate LibreChat into your language, we'd love your contribution! Improving our translations not only makes LibreChat more accessible to users around the world but also enhances the overall user experience. Please check out our [Translation Guide](https://www.librechat.ai/docs/translation).
 
 ---
 
-## Star History
+## 💖 This project exists in its current state thanks to all the people who contribute
 
-<a href="https://star-history.com/#danny-avila/LibreChat&Date">
-  <img alt="Star History Chart" src="https://api.star-history.com/svg?repos=danny-avila/LibreChat&type=Date&theme=dark" onerror="this.src='https://api.star-history.com/svg?repos=danny-avila/LibreChat&type=Date'" />
-</a>
-
----
-
-## Sponsors
-
-  Sponsored by <a href="https://github.com/mjtechguy"><b>@mjtechguy</b></a>, <a href="https://github.com/SphaeroX"><b>@SphaeroX</b></a>, <a href="https://github.com/DavidDev1334"><b>@DavidDev1334</b></a>, <a href="https://github.com/fuegovic"><b>@fuegovic</b></a>, <a href="https://github.com/Pharrcyde"><b>@Pharrcyde</b></a> 
-  
----
-
-## Contributors
-Contributions and suggestions bug reports and fixes are welcome!
-Please read the documentation before you do!
-
----
-
-For new features, components, or extensions, please open an issue and discuss before sending a PR. 
-
-- Join the [Discord community](https://discord.gg/uDyZ5Tzhct)
-
-This project exists in its current state thanks to all the people who contribute
----
 <a href="https://github.com/danny-avila/LibreChat/graphs/contributors">
   <img src="https://contrib.rocks/image?repo=danny-avila/LibreChat" />
 </a>
+
+---
+
+## 🎉 Special Thanks
+
+We thank [Locize](https://locize.com) for their translation management tools that support multiple languages in LibreChat.
+
+<p align="center">
+  <a href="https://locize.com" target="_blank" rel="noopener noreferrer">
+    <img src="https://locize.com/img/locize_color.svg" alt="Locize Logo" height="50">
+  </a>
+</p>

api/app/bingai.js

@@ -1,113 +0,0 @@
-require('dotenv').config();
-const { KeyvFile } = require('keyv-file');
-const { getUserKey, checkUserKeyExpiry } = require('../server/services/UserService');
-
-const askBing = async ({
-  text,
-  parentMessageId,
-  conversationId,
-  jailbreak,
-  jailbreakConversationId,
-  context,
-  systemMessage,
-  conversationSignature,
-  clientId,
-  invocationId,
-  toneStyle,
-  key: expiresAt,
-  onProgress,
-  userId,
-}) => {
-  const isUserProvided = process.env.BINGAI_TOKEN === 'user_provided';
-
-  let key = null;
-  if (expiresAt && isUserProvided) {
-    checkUserKeyExpiry(
-      expiresAt,
-      'Your BingAI Cookies have expired. Please provide your cookies again.',
-    );
-    key = await getUserKey({ userId, name: 'bingAI' });
-  }
-
-  const { BingAIClient } = await import('nodejs-gpt');
-  const store = {
-    store: new KeyvFile({ filename: './data/cache.json' }),
-  };
-
-  const bingAIClient = new BingAIClient({
-    // "_U" cookie from bing.com
-    // userToken:
-    //   isUserProvided ? key : process.env.BINGAI_TOKEN ?? null,
-    // If the above doesn't work, provide all your cookies as a string instead
-    cookies: isUserProvided ? key : process.env.BINGAI_TOKEN ?? null,
-    debug: false,
-    cache: store,
-    host: process.env.BINGAI_HOST || null,
-    proxy: process.env.PROXY || null,
-  });
-
-  let options = {};
-
-  if (jailbreakConversationId == 'false') {
-    jailbreakConversationId = false;
-  }
-
-  if (jailbreak) {
-    options = {
-      jailbreakConversationId: jailbreakConversationId || jailbreak,
-      context,
-      systemMessage,
-      parentMessageId,
-      toneStyle,
-      onProgress,
-      clientOptions: {
-        features: {
-          genImage: {
-            server: {
-              enable: true,
-              type: 'markdown_list',
-            },
-          },
-        },
-      },
-    };
-  } else {
-    options = {
-      conversationId,
-      context,
-      systemMessage,
-      parentMessageId,
-      toneStyle,
-      onProgress,
-      clientOptions: {
-        features: {
-          genImage: {
-            server: {
-              enable: true,
-              type: 'markdown_list',
-            },
-          },
-        },
-      },
-    };
-
-    // don't give those parameters for new conversation
-    // for new conversation, conversationSignature always is null
-    if (conversationSignature) {
-      options.encryptedConversationSignature = conversationSignature;
-      options.clientId = clientId;
-      options.invocationId = invocationId;
-    }
-  }
-
-  console.log('bing options', options);
-
-  const res = await bingAIClient.sendMessage(text, options);
-
-  return res;
-
-  // for reference:
-  // https://github.com/waylaidwanderer/node-chatgpt-api/blob/main/demos/use-bing-client.js
-};
-
-module.exports = { askBing };

api/app/chatgpt-browser.js

@@ -1,59 +0,0 @@
-require('dotenv').config();
-const { KeyvFile } = require('keyv-file');
-const { getUserKey, checkUserKeyExpiry } = require('../server/services/UserService');
-
-const browserClient = async ({
-  text,
-  parentMessageId,
-  conversationId,
-  model,
-  key: expiresAt,
-  onProgress,
-  onEventMessage,
-  abortController,
-  userId,
-}) => {
-  const isUserProvided = process.env.CHATGPT_TOKEN === 'user_provided';
-
-  let key = null;
-  if (expiresAt && isUserProvided) {
-    checkUserKeyExpiry(
-      expiresAt,
-      'Your ChatGPT Access Token has expired. Please provide your token again.',
-    );
-    key = await getUserKey({ userId, name: 'chatGPTBrowser' });
-  }
-
-  const { ChatGPTBrowserClient } = await import('nodejs-gpt');
-  const store = {
-    store: new KeyvFile({ filename: './data/cache.json' }),
-  };
-
-  const clientOptions = {
-    // Warning: This will expose your access token to a third party. Consider the risks before using this.
-    reverseProxyUrl:
-      process.env.CHATGPT_REVERSE_PROXY ?? 'https://ai.fakeopen.com/api/conversation',
-    // Access token from https://chat.openai.com/api/auth/session
-    accessToken: isUserProvided ? key : process.env.CHATGPT_TOKEN ?? null,
-    model: model,
-    debug: false,
-    proxy: process.env.PROXY ?? null,
-    user: userId,
-  };
-
-  const client = new ChatGPTBrowserClient(clientOptions, store);
-  let options = { onProgress, onEventMessage, abortController };
-
-  if (!!parentMessageId && !!conversationId) {
-    options = { ...options, parentMessageId, conversationId };
-  }
-
-  if (parentMessageId === '00000000-0000-0000-0000-000000000000') {
-    delete options.conversationId;
-  }
-
-  const res = await client.sendMessage(text, options);
-  return res;
-};
-
-module.exports = { browserClient };

api/app/clients/AnthropicClient.js

@@ -1,21 +1,90 @@
-// const { Agent, ProxyAgent } = require('undici');
-const BaseClient = require('./BaseClient');
-const { encoding_for_model: encodingForModel, get_encoding: getEncoding } = require('tiktoken');
 const Anthropic = require('@anthropic-ai/sdk');
+const { HttpsProxyAgent } = require('https-proxy-agent');
+const {
+  Constants,
+  EModelEndpoint,
+  anthropicSettings,
+  getResponseSender,
+  validateVisionModel,
+} = require('librechat-data-provider');
+const { SplitStreamHandler: _Handler, GraphEvents } = require('@librechat/agents');
+const {
+  truncateText,
+  formatMessage,
+  addCacheControl,
+  titleFunctionPrompt,
+  parseParamFromPrompt,
+  createContextHandlers,
+} = require('./prompts');
+const {
+  getClaudeHeaders,
+  configureReasoning,
+  checkPromptCacheSupport,
+} = require('~/server/services/Endpoints/anthropic/helpers');
+const { getModelMaxTokens, getModelMaxOutputTokens, matchModelName } = require('~/utils');
+const { spendTokens, spendStructuredTokens } = require('~/models/spendTokens');
+const { encodeAndFormat } = require('~/server/services/Files/images/encode');
+const Tokenizer = require('~/server/services/Tokenizer');
+const { logger, sendEvent } = require('~/config');
+const { sleep } = require('~/server/utils');
+const BaseClient = require('./BaseClient');
 
 const HUMAN_PROMPT = '\n\nHuman:';
 const AI_PROMPT = '\n\nAssistant:';
 
-const tokenizersCache = {};
+class SplitStreamHandler extends _Handler {
+  getDeltaContent(chunk) {
+    return (chunk?.delta?.text ?? chunk?.completion) || '';
+  }
+  getReasoningDelta(chunk) {
+    return chunk?.delta?.thinking || '';
+  }
+}
+
+/** Helper function to introduce a delay before retrying */
+function delayBeforeRetry(attempts, baseDelay = 1000) {
+  return new Promise((resolve) => setTimeout(resolve, baseDelay * attempts));
+}
+
+const tokenEventTypes = new Set(['message_start', 'message_delta']);
+const { legacy } = anthropicSettings;
 
 class AnthropicClient extends BaseClient {
-  constructor(apiKey, options = {}, cacheOptions = {}) {
-    super(apiKey, options, cacheOptions);
+  constructor(apiKey, options = {}) {
+    super(apiKey, options);
     this.apiKey = apiKey || process.env.ANTHROPIC_API_KEY;
-    this.sender = 'Anthropic';
     this.userLabel = HUMAN_PROMPT;
     this.assistantLabel = AI_PROMPT;
+    this.contextStrategy = options.contextStrategy
+      ? options.contextStrategy.toLowerCase()
+      : 'discard';
     this.setOptions(options);
+    /** @type {string | undefined} */
+    this.systemMessage;
+    /** @type {AnthropicMessageStartEvent| undefined} */
+    this.message_start;
+    /** @type {AnthropicMessageDeltaEvent| undefined} */
+    this.message_delta;
+    /** Whether the model is part of the Claude 3 Family
+     * @type {boolean} */
+    this.isClaude3;
+    /** Whether to use Messages API or Completions API
+     * @type {boolean} */
+    this.useMessages;
+    /** Whether or not the model is limited to the legacy amount of output tokens
+     * @type {boolean} */
+    this.isLegacyOutput;
+    /** Whether or not the model supports Prompt Caching
+     * @type {boolean} */
+    this.supportsCacheControl;
+    /** The key for the usage object's input tokens
+     * @type {string} */
+    this.inputTokensKey = 'input_tokens';
+    /** The key for the usage object's output tokens
+     * @type {string} */
+    this.outputTokensKey = 'output_tokens';
+    /** @type {SplitStreamHandler | undefined} */
+    this.streamHandler;
   }
 
   setOptions(options) {
@@ -35,19 +104,46 @@ class AnthropicClient extends BaseClient {
       this.options = options;
     }
 
-    const modelOptions = this.options.modelOptions || {};
-    this.modelOptions = {
-      ...modelOptions,
-      // set some good defaults (check for undefined in some cases because they may be 0)
-      model: modelOptions.model || 'claude-1',
-      temperature: typeof modelOptions.temperature === 'undefined' ? 0.7 : modelOptions.temperature, // 0 - 1, 0.7 is recommended
-      topP: typeof modelOptions.topP === 'undefined' ? 0.7 : modelOptions.topP, // 0 - 1, default: 0.7
-      topK: typeof modelOptions.topK === 'undefined' ? 40 : modelOptions.topK, // 1-40, default: 40
-      stop: modelOptions.stop, // no stop method for now
-    };
+    this.modelOptions = Object.assign(
+      {
+        model: anthropicSettings.model.default,
+      },
+      this.modelOptions,
+      this.options.modelOptions,
+    );
 
-    this.maxContextTokens = this.options.maxContextTokens || 99999;
-    this.maxResponseTokens = this.modelOptions.maxOutputTokens || 1500;
+    const modelMatch = matchModelName(this.modelOptions.model, EModelEndpoint.anthropic);
+    this.isClaude3 = modelMatch.includes('claude-3');
+    this.isLegacyOutput = !(
+      /claude-3[-.]5-sonnet/.test(modelMatch) || /claude-3[-.]7/.test(modelMatch)
+    );
+    this.supportsCacheControl = this.options.promptCache && checkPromptCacheSupport(modelMatch);
+
+    if (
+      this.isLegacyOutput &&
+      this.modelOptions.maxOutputTokens &&
+      this.modelOptions.maxOutputTokens > legacy.maxOutputTokens.default
+    ) {
+      this.modelOptions.maxOutputTokens = legacy.maxOutputTokens.default;
+    }
+
+    this.useMessages = this.isClaude3 || !!this.options.attachments;
+
+    this.defaultVisionModel = this.options.visionModel ?? 'claude-3-sonnet-20240229';
+    this.options.attachments?.then((attachments) => this.checkVisionRequest(attachments));
+
+    this.maxContextTokens =
+      this.options.maxContextTokens ??
+      getModelMaxTokens(this.modelOptions.model, EModelEndpoint.anthropic) ??
+      100000;
+    this.maxResponseTokens =
+      this.modelOptions.maxOutputTokens ??
+      getModelMaxOutputTokens(
+        this.modelOptions.model,
+        this.options.endpointType ?? this.options.endpoint,
+        this.options.endpointTokenConfig,
+      ) ??
+      anthropicSettings.maxOutputTokens.reset(this.modelOptions.model);
     this.maxPromptTokens =
       this.options.maxPromptTokens || this.maxContextTokens - this.maxResponseTokens;
 
@@ -59,35 +155,193 @@ class AnthropicClient extends BaseClient {
       );
     }
 
+    this.sender =
+      this.options.sender ??
+      getResponseSender({
+        model: this.modelOptions.model,
+        endpoint: EModelEndpoint.anthropic,
+        modelLabel: this.options.modelLabel,
+      });
+
     this.startToken = '||>';
     this.endToken = '';
-    this.gptEncoder = this.constructor.getTokenizer('cl100k_base');
-
-    if (!this.modelOptions.stop) {
-      const stopTokens = [this.startToken];
-      if (this.endToken && this.endToken !== this.startToken) {
-        stopTokens.push(this.endToken);
-      }
-      stopTokens.push(`${this.userLabel}`);
-      stopTokens.push('<|diff_marker|>');
-
-      this.modelOptions.stop = stopTokens;
-    }
 
     return this;
   }
 
-  getClient() {
-    if (this.options.reverseProxyUrl) {
-      return new Anthropic({
-        apiKey: this.apiKey,
-        baseURL: this.options.reverseProxyUrl,
-      });
-    } else {
-      return new Anthropic({
-        apiKey: this.apiKey,
-      });
+  /**
+   * Get the initialized Anthropic client.
+   * @param {Partial<Anthropic.ClientOptions>} requestOptions - The options for the client.
+   * @returns {Anthropic} The Anthropic client instance.
+   */
+  getClient(requestOptions) {
+    /** @type {Anthropic.ClientOptions} */
+    const options = {
+      fetch: this.fetch,
+      apiKey: this.apiKey,
+    };
+
+    if (this.options.proxy) {
+      options.httpAgent = new HttpsProxyAgent(this.options.proxy);
     }
+
+    if (this.options.reverseProxyUrl) {
+      options.baseURL = this.options.reverseProxyUrl;
+    }
+
+    const headers = getClaudeHeaders(requestOptions?.model, this.supportsCacheControl);
+    if (headers) {
+      options.defaultHeaders = headers;
+    }
+
+    return new Anthropic(options);
+  }
+
+  /**
+   * Get stream usage as returned by this client's API response.
+   * @returns {AnthropicStreamUsage} The stream usage object.
+   */
+  getStreamUsage() {
+    const inputUsage = this.message_start?.message?.usage ?? {};
+    const outputUsage = this.message_delta?.usage ?? {};
+    return Object.assign({}, inputUsage, outputUsage);
+  }
+
+  /**
+   * Calculates the correct token count for the current user message based on the token count map and API usage.
+   * Edge case: If the calculation results in a negative value, it returns the original estimate.
+   * If revisiting a conversation with a chat history entirely composed of token estimates,
+   * the cumulative token count going forward should become more accurate as the conversation progresses.
+   * @param {Object} params - The parameters for the calculation.
+   * @param {Record<string, number>} params.tokenCountMap - A map of message IDs to their token counts.
+   * @param {string} params.currentMessageId - The ID of the current message to calculate.
+   * @param {AnthropicStreamUsage} params.usage - The usage object returned by the API.
+   * @returns {number} The correct token count for the current user message.
+   */
+  calculateCurrentTokenCount({ tokenCountMap, currentMessageId, usage }) {
+    const originalEstimate = tokenCountMap[currentMessageId] || 0;
+
+    if (!usage || typeof usage.input_tokens !== 'number') {
+      return originalEstimate;
+    }
+
+    tokenCountMap[currentMessageId] = 0;
+    const totalTokensFromMap = Object.values(tokenCountMap).reduce((sum, count) => {
+      const numCount = Number(count);
+      return sum + (isNaN(numCount) ? 0 : numCount);
+    }, 0);
+    const totalInputTokens =
+      (usage.input_tokens ?? 0) +
+      (usage.cache_creation_input_tokens ?? 0) +
+      (usage.cache_read_input_tokens ?? 0);
+
+    const currentMessageTokens = totalInputTokens - totalTokensFromMap;
+    return currentMessageTokens > 0 ? currentMessageTokens : originalEstimate;
+  }
+
+  /**
+   * Get Token Count for LibreChat Message
+   * @param {TMessage} responseMessage
+   * @returns {number}
+   */
+  getTokenCountForResponse(responseMessage) {
+    return this.getTokenCountForMessage({
+      role: 'assistant',
+      content: responseMessage.text,
+    });
+  }
+
+  /**
+   *
+   * Checks if the model is a vision model based on request attachments and sets the appropriate options:
+   * - Sets `this.modelOptions.model` to `gpt-4-vision-preview` if the request is a vision request.
+   * - Sets `this.isVisionModel` to `true` if vision request.
+   * - Deletes `this.modelOptions.stop` if vision request.
+   * @param {MongoFile[]} attachments
+   */
+  checkVisionRequest(attachments) {
+    const availableModels = this.options.modelsConfig?.[EModelEndpoint.anthropic];
+    this.isVisionModel = validateVisionModel({ model: this.modelOptions.model, availableModels });
+
+    const visionModelAvailable = availableModels?.includes(this.defaultVisionModel);
+    if (
+      attachments &&
+      attachments.some((file) => file?.type && file?.type?.includes('image')) &&
+      visionModelAvailable &&
+      !this.isVisionModel
+    ) {
+      this.modelOptions.model = this.defaultVisionModel;
+      this.isVisionModel = true;
+    }
+  }
+
+  /**
+   * Calculate the token cost in tokens for an image based on its dimensions and detail level.
+   *
+   * For reference, see: https://docs.anthropic.com/claude/docs/vision#image-costs
+   *
+   * @param {Object} image - The image object.
+   * @param {number} image.width - The width of the image.
+   * @param {number} image.height - The height of the image.
+   * @returns {number} The calculated token cost measured by tokens.
+   *
+   */
+  calculateImageTokenCost({ width, height }) {
+    return Math.ceil((width * height) / 750);
+  }
+
+  async addImageURLs(message, attachments) {
+    const { files, image_urls } = await encodeAndFormat(
+      this.options.req,
+      attachments,
+      EModelEndpoint.anthropic,
+    );
+    message.image_urls = image_urls.length ? image_urls : undefined;
+    return files;
+  }
+
+  /**
+   * @param {object} params
+   * @param {number} params.promptTokens
+   * @param {number} params.completionTokens
+   * @param {AnthropicStreamUsage} [params.usage]
+   * @param {string} [params.model]
+   * @param {string} [params.context='message']
+   * @returns {Promise<void>}
+   */
+  async recordTokenUsage({ promptTokens, completionTokens, usage, model, context = 'message' }) {
+    if (usage != null && usage?.input_tokens != null) {
+      const input = usage.input_tokens ?? 0;
+      const write = usage.cache_creation_input_tokens ?? 0;
+      const read = usage.cache_read_input_tokens ?? 0;
+
+      await spendStructuredTokens(
+        {
+          context,
+          user: this.user,
+          conversationId: this.conversationId,
+          model: model ?? this.modelOptions.model,
+          endpointTokenConfig: this.options.endpointTokenConfig,
+        },
+        {
+          promptTokens: { input, write, read },
+          completionTokens,
+        },
+      );
+
+      return;
+    }
+
+    await spendTokens(
+      {
+        context,
+        user: this.user,
+        conversationId: this.conversationId,
+        model: model ?? this.modelOptions.model,
+        endpointTokenConfig: this.options.endpointTokenConfig,
+      },
+      { promptTokens, completionTokens },
+    );
   }
 
   async buildMessages(messages, parentMessageId) {
@@ -95,32 +349,148 @@ class AnthropicClient extends BaseClient {
       messages,
       parentMessageId,
     });
-    if (this.options.debug) {
-      console.debug('AnthropicClient: orderedMessages', orderedMessages, parentMessageId);
+
+    logger.debug('[AnthropicClient] orderedMessages', { orderedMessages, parentMessageId });
+
+    if (this.options.attachments) {
+      const attachments = await this.options.attachments;
+      const images = attachments.filter((file) => file.type.includes('image'));
+
+      if (images.length && !this.isVisionModel) {
+        throw new Error('Images are only supported with the Claude 3 family of models');
+      }
+
+      const latestMessage = orderedMessages[orderedMessages.length - 1];
+
+      if (this.message_file_map) {
+        this.message_file_map[latestMessage.messageId] = attachments;
+      } else {
+        this.message_file_map = {
+          [latestMessage.messageId]: attachments,
+        };
+      }
+
+      const files = await this.addImageURLs(latestMessage, attachments);
+
+      this.options.attachments = files;
     }
 
-    const formattedMessages = orderedMessages.map((message) => ({
-      author: message.isCreatedByUser ? this.userLabel : this.assistantLabel,
-      content: message?.content ?? message.text,
-    }));
+    if (this.message_file_map) {
+      this.contextHandlers = createContextHandlers(
+        this.options.req,
+        orderedMessages[orderedMessages.length - 1].text,
+      );
+    }
+
+    const formattedMessages = orderedMessages.map((message, i) => {
+      const formattedMessage = this.useMessages
+        ? formatMessage({
+          message,
+          endpoint: EModelEndpoint.anthropic,
+        })
+        : {
+          author: message.isCreatedByUser ? this.userLabel : this.assistantLabel,
+          content: message?.content ?? message.text,
+        };
+
+      const needsTokenCount = this.contextStrategy && !orderedMessages[i].tokenCount;
+      /* If tokens were never counted, or, is a Vision request and the message has files, count again */
+      if (needsTokenCount || (this.isVisionModel && (message.image_urls || message.files))) {
+        orderedMessages[i].tokenCount = this.getTokenCountForMessage(formattedMessage);
+      }
+
+      /* If message has files, calculate image token cost */
+      if (this.message_file_map && this.message_file_map[message.messageId]) {
+        const attachments = this.message_file_map[message.messageId];
+        for (const file of attachments) {
+          if (file.embedded) {
+            this.contextHandlers?.processFile(file);
+            continue;
+          }
+
+          orderedMessages[i].tokenCount += this.calculateImageTokenCost({
+            width: file.width,
+            height: file.height,
+          });
+        }
+      }
+
+      formattedMessage.tokenCount = orderedMessages[i].tokenCount;
+      return formattedMessage;
+    });
+
+    if (this.contextHandlers) {
+      this.augmentedPrompt = await this.contextHandlers.createContext();
+      this.options.promptPrefix = this.augmentedPrompt + (this.options.promptPrefix ?? '');
+    }
+
+    let { context: messagesInWindow, remainingContextTokens } =
+      await this.getMessagesWithinTokenLimit({ messages: formattedMessages });
+
+    const tokenCountMap = orderedMessages
+      .slice(orderedMessages.length - messagesInWindow.length)
+      .reduce((map, message, index) => {
+        const { messageId } = message;
+        if (!messageId) {
+          return map;
+        }
+
+        map[messageId] = orderedMessages[index].tokenCount;
+        return map;
+      }, {});
+
+    logger.debug('[AnthropicClient]', {
+      messagesInWindow: messagesInWindow.length,
+      remainingContextTokens,
+    });
 
     let lastAuthor = '';
     let groupedMessages = [];
 
-    for (let message of formattedMessages) {
+    for (let i = 0; i < messagesInWindow.length; i++) {
+      const message = messagesInWindow[i];
+      const author = message.role ?? message.author;
       // If last author is not same as current author, add to new group
-      if (lastAuthor !== message.author) {
-        groupedMessages.push({
-          author: message.author,
+      if (lastAuthor !== author) {
+        const newMessage = {
           content: [message.content],
-        });
-        lastAuthor = message.author;
+        };
+
+        if (message.role) {
+          newMessage.role = message.role;
+        } else {
+          newMessage.author = message.author;
+        }
+
+        groupedMessages.push(newMessage);
+        lastAuthor = author;
         // If same author, append content to the last group
       } else {
         groupedMessages[groupedMessages.length - 1].content.push(message.content);
       }
     }
 
+    groupedMessages = groupedMessages.map((msg, i) => {
+      const isLast = i === groupedMessages.length - 1;
+      if (msg.content.length === 1) {
+        const content = msg.content[0];
+        return {
+          ...msg,
+          // reason: final assistant content cannot end with trailing whitespace
+          content:
+            isLast && this.useMessages && msg.role === 'assistant' && typeof content === 'string'
+              ? content?.trim()
+              : content,
+        };
+      }
+
+      if (!this.useMessages && msg.tokenCount) {
+        delete msg.tokenCount;
+      }
+
+      return msg;
+    });
+
     let identityPrefix = '';
     if (this.options.userLabel) {
       identityPrefix = `\nHuman's name: ${this.options.userLabel}`;
@@ -130,7 +500,10 @@ class AnthropicClient extends BaseClient {
       identityPrefix = `${identityPrefix}\nYou are ${this.options.modelLabel}`;
     }
 
-    let promptPrefix = (this.options.promptPrefix || '').trim();
+    let promptPrefix = (this.options.promptPrefix ?? '').trim();
+    if (typeof this.options.artifactsPrompt === 'string' && this.options.artifactsPrompt) {
+      promptPrefix = `${promptPrefix ?? ''}\n${this.options.artifactsPrompt}`.trim();
+    }
     if (promptPrefix) {
       // If the prompt prefix doesn't end with the end token, add it.
       if (!promptPrefix.endsWith(`${this.endToken}`)) {
@@ -146,9 +519,10 @@ class AnthropicClient extends BaseClient {
     // Prompt AI to respond, empty if last message was from AI
     let isEdited = lastAuthor === this.assistantLabel;
     const promptSuffix = isEdited ? '' : `${promptPrefix}${this.assistantLabel}\n`;
-    let currentTokenCount = isEdited
-      ? this.getTokenCount(promptPrefix)
-      : this.getTokenCount(promptSuffix);
+    let currentTokenCount =
+      isEdited || this.useMessages
+        ? this.getTokenCount(promptPrefix)
+        : this.getTokenCount(promptSuffix);
 
     let promptBody = '';
     const maxTokenCount = this.maxPromptTokens;
@@ -216,7 +590,69 @@ class AnthropicClient extends BaseClient {
       return true;
     };
 
-    await buildPromptBody();
+    const messagesPayload = [];
+    const buildMessagesPayload = async () => {
+      let canContinue = true;
+
+      if (promptPrefix) {
+        this.systemMessage = promptPrefix;
+      }
+
+      while (currentTokenCount < maxTokenCount && groupedMessages.length > 0 && canContinue) {
+        const message = groupedMessages.pop();
+
+        let tokenCountForMessage = message.tokenCount ?? this.getTokenCountForMessage(message);
+
+        const newTokenCount = currentTokenCount + tokenCountForMessage;
+        const exceededMaxCount = newTokenCount > maxTokenCount;
+
+        if (exceededMaxCount && messagesPayload.length === 0) {
+          throw new Error(
+            `Prompt is too long. Max token count is ${maxTokenCount}, but prompt is ${newTokenCount} tokens long.`,
+          );
+        } else if (exceededMaxCount) {
+          canContinue = false;
+          break;
+        }
+
+        delete message.tokenCount;
+        messagesPayload.unshift(message);
+        currentTokenCount = newTokenCount;
+
+        // Switch off isEdited after using it once
+        if (isEdited && message.role === 'assistant') {
+          isEdited = false;
+        }
+
+        // Wait for next tick to avoid blocking the event loop
+        await new Promise((resolve) => setImmediate(resolve));
+      }
+    };
+
+    const processTokens = () => {
+      // Add 2 tokens for metadata after all messages have been counted.
+      currentTokenCount += 2;
+
+      // Use up to `this.maxContextTokens` tokens (prompt + response), but try to leave `this.maxTokens` tokens for the response.
+      this.modelOptions.maxOutputTokens = Math.min(
+        this.maxContextTokens - currentTokenCount,
+        this.maxResponseTokens,
+      );
+    };
+
+    if (this.modelOptions.model.includes('claude-3')) {
+      await buildMessagesPayload();
+      processTokens();
+      return {
+        prompt: messagesPayload,
+        context: messagesInWindow,
+        promptTokens: currentTokenCount,
+        tokenCountMap,
+      };
+    } else {
+      await buildPromptBody();
+      processTokens();
+    }
 
     if (nextMessage.remove) {
       promptBody = promptBody.replace(nextMessage.messageString, '');
@@ -226,20 +662,53 @@ class AnthropicClient extends BaseClient {
 
     let prompt = `${promptBody}${promptSuffix}`;
 
-    // Add 2 tokens for metadata after all messages have been counted.
-    currentTokenCount += 2;
-
-    // Use up to `this.maxContextTokens` tokens (prompt + response), but try to leave `this.maxTokens` tokens for the response.
-    this.modelOptions.maxOutputTokens = Math.min(
-      this.maxContextTokens - currentTokenCount,
-      this.maxResponseTokens,
-    );
-
-    return { prompt, context };
+    return { prompt, context, promptTokens: currentTokenCount, tokenCountMap };
   }
 
   getCompletion() {
-    console.log('AnthropicClient doesn\'t use getCompletion (all handled in sendCompletion)');
+    logger.debug('AnthropicClient doesn\'t use getCompletion (all handled in sendCompletion)');
+  }
+
+  /**
+   * Creates a message or completion response using the Anthropic client.
+   * @param {Anthropic} client - The Anthropic client instance.
+   * @param {Anthropic.default.MessageCreateParams | Anthropic.default.CompletionCreateParams} options - The options for the message or completion.
+   * @param {boolean} useMessages - Whether to use messages or completions. Defaults to `this.useMessages`.
+   * @returns {Promise<Anthropic.default.Message | Anthropic.default.Completion>} The response from the Anthropic client.
+   */
+  async createResponse(client, options, useMessages) {
+    return (useMessages ?? this.useMessages)
+      ? await client.messages.create(options)
+      : await client.completions.create(options);
+  }
+
+  getMessageMapMethod() {
+    /**
+     * @param {TMessage} msg
+     */
+    return (msg) => {
+      if (msg.text != null && msg.text && msg.text.startsWith(':::thinking')) {
+        msg.text = msg.text.replace(/:::thinking.*?:::/gs, '').trim();
+      }
+
+      return msg;
+    };
+  }
+
+  /**
+   * @param {string[]} [intermediateReply]
+   * @returns {string}
+   */
+  getStreamText(intermediateReply) {
+    if (!this.streamHandler) {
+      return intermediateReply?.join('') ?? '';
+    }
+
+    const reasoningText = this.streamHandler.reasoningTokens.join('');
+
+    const reasoningBlock = reasoningText.length > 0 ? `:::thinking\n${reasoningText}\n:::\n` : '';
+
+    return `${reasoningBlock}${this.streamHandler.tokens.join('')}`;
   }
 
   async sendCompletion(payload, { onProgress, abortController }) {
@@ -254,19 +723,11 @@ class AnthropicClient extends BaseClient {
       modelOptions.stream = true;
     }
 
-    const { debug } = this.options;
-    if (debug) {
-      console.debug();
-      console.debug(modelOptions);
-      console.debug();
-    }
-
-    const client = this.getClient();
+    logger.debug('modelOptions', { modelOptions });
     const metadata = {
       user_id: this.user,
     };
 
-    let text = '';
     const {
       stream,
       model,
@@ -276,79 +737,237 @@ class AnthropicClient extends BaseClient {
       topP: top_p,
       topK: top_k,
     } = this.modelOptions;
-    const requestOptions = {
-      prompt: payload,
+
+    let requestOptions = {
       model,
       stream: stream || true,
-      max_tokens_to_sample: maxOutputTokens || 1500,
       stop_sequences,
       temperature,
       metadata,
-      top_p,
-      top_k,
     };
-    if (this.options.debug) {
-      console.log('AnthropicClient: requestOptions');
-      console.dir(requestOptions, { depth: null });
-    }
-    const response = await client.completions.create(requestOptions);
 
-    signal.addEventListener('abort', () => {
-      if (this.options.debug) {
-        console.log('AnthropicClient: message aborted!');
+    if (!/claude-3[-.]7/.test(model)) {
+      if (top_p !== undefined) {
+        requestOptions.top_p = top_p;
       }
-      response.controller.abort();
-    });
-
-    for await (const completion of response) {
-      if (this.options.debug) {
-        // Uncomment to debug message stream
-        // console.debug(completion);
+      if (top_k !== undefined) {
+        requestOptions.top_k = top_k;
       }
-      text += completion.completion;
-      onProgress(completion.completion);
     }
 
-    signal.removeEventListener('abort', () => {
-      if (this.options.debug) {
-        console.log('AnthropicClient: message aborted!');
-      }
-      response.controller.abort();
+    if (this.useMessages) {
+      requestOptions.messages = payload;
+      requestOptions.max_tokens =
+        maxOutputTokens || anthropicSettings.maxOutputTokens.reset(requestOptions.model);
+    } else {
+      requestOptions.prompt = payload;
+      requestOptions.max_tokens_to_sample = maxOutputTokens || legacy.maxOutputTokens.default;
+    }
+
+    requestOptions = configureReasoning(requestOptions, {
+      thinking: this.options.thinking,
+      thinkingBudget: this.options.thinkingBudget,
     });
 
-    return text.trim();
+    if (this.systemMessage && this.supportsCacheControl === true) {
+      requestOptions.system = [
+        {
+          type: 'text',
+          text: this.systemMessage,
+          cache_control: { type: 'ephemeral' },
+        },
+      ];
+    } else if (this.systemMessage) {
+      requestOptions.system = this.systemMessage;
+    }
+
+    if (this.supportsCacheControl === true && this.useMessages) {
+      requestOptions.messages = addCacheControl(requestOptions.messages);
+    }
+
+    logger.debug('[AnthropicClient]', { ...requestOptions });
+    this.streamHandler = new SplitStreamHandler({
+      accumulate: true,
+      runId: this.responseMessageId,
+      handlers: {
+        [GraphEvents.ON_RUN_STEP]: (event) => sendEvent(this.options.res, event),
+        [GraphEvents.ON_MESSAGE_DELTA]: (event) => sendEvent(this.options.res, event),
+        [GraphEvents.ON_REASONING_DELTA]: (event) => sendEvent(this.options.res, event),
+      },
+    });
+
+    let intermediateReply = this.streamHandler.tokens;
+
+    const maxRetries = 3;
+    const streamRate = this.options.streamRate ?? Constants.DEFAULT_STREAM_RATE;
+    async function processResponse() {
+      let attempts = 0;
+
+      while (attempts < maxRetries) {
+        let response;
+        try {
+          const client = this.getClient(requestOptions);
+          response = await this.createResponse(client, requestOptions);
+
+          signal.addEventListener('abort', () => {
+            logger.debug('[AnthropicClient] message aborted!');
+            if (response.controller?.abort) {
+              response.controller.abort();
+            }
+          });
+
+          for await (const completion of response) {
+            const type = completion?.type ?? '';
+            if (tokenEventTypes.has(type)) {
+              logger.debug(`[AnthropicClient] ${type}`, completion);
+              this[type] = completion;
+            }
+            this.streamHandler.handle(completion);
+            await sleep(streamRate);
+          }
+
+          break;
+        } catch (error) {
+          attempts += 1;
+          logger.warn(
+            `User: ${this.user} | Anthropic Request ${attempts} failed: ${error.message}`,
+          );
+
+          if (attempts < maxRetries) {
+            await delayBeforeRetry(attempts, 350);
+          } else if (this.streamHandler && this.streamHandler.reasoningTokens.length) {
+            return this.getStreamText();
+          } else if (intermediateReply.length > 0) {
+            return this.getStreamText(intermediateReply);
+          } else {
+            throw new Error(`Operation failed after ${maxRetries} attempts: ${error.message}`);
+          }
+        } finally {
+          signal.removeEventListener('abort', () => {
+            logger.debug('[AnthropicClient] message aborted!');
+            if (response.controller?.abort) {
+              response.controller.abort();
+            }
+          });
+        }
+      }
+    }
+
+    await processResponse.bind(this)();
+    return this.getStreamText(intermediateReply);
   }
 
   getSaveOptions() {
     return {
+      maxContextTokens: this.options.maxContextTokens,
+      artifacts: this.options.artifacts,
       promptPrefix: this.options.promptPrefix,
       modelLabel: this.options.modelLabel,
+      promptCache: this.options.promptCache,
+      thinking: this.options.thinking,
+      thinkingBudget: this.options.thinkingBudget,
+      resendFiles: this.options.resendFiles,
+      iconURL: this.options.iconURL,
+      greeting: this.options.greeting,
+      spec: this.options.spec,
       ...this.modelOptions,
     };
   }
 
   getBuildMessagesOptions() {
-    if (this.options.debug) {
-      console.log('AnthropicClient doesn\'t use getBuildMessagesOptions');
-    }
+    logger.debug('AnthropicClient doesn\'t use getBuildMessagesOptions');
   }
 
-  static getTokenizer(encoding, isModelName = false, extendSpecialTokens = {}) {
-    if (tokenizersCache[encoding]) {
-      return tokenizersCache[encoding];
-    }
-    let tokenizer;
-    if (isModelName) {
-      tokenizer = encodingForModel(encoding, extendSpecialTokens);
-    } else {
-      tokenizer = getEncoding(encoding, extendSpecialTokens);
-    }
-    tokenizersCache[encoding] = tokenizer;
-    return tokenizer;
+  getEncoding() {
+    return 'cl100k_base';
   }
 
+  /**
+   * Returns the token count of a given text. It also checks and resets the tokenizers if necessary.
+   * @param {string} text - The text to get the token count for.
+   * @returns {number} The token count of the given text.
+   */
   getTokenCount(text) {
-    return this.gptEncoder.encode(text, 'all').length;
+    const encoding = this.getEncoding();
+    return Tokenizer.getTokenCount(text, encoding);
+  }
+
+  /**
+   * Generates a concise title for a conversation based on the user's input text and response.
+   * Involves sending a chat completion request with specific instructions for title generation.
+   *
+   * This function capitlizes on [Anthropic's function calling training](https://docs.anthropic.com/claude/docs/functions-external-tools).
+   *
+   * @param {Object} params - The parameters for the conversation title generation.
+   * @param {string} params.text - The user's input.
+   * @param {string} [params.responseText=''] - The AI's immediate response to the user.
+   *
+   * @returns {Promise<string | 'New Chat'>} A promise that resolves to the generated conversation title.
+   *                            In case of failure, it will return the default title, "New Chat".
+   */
+  async titleConvo({ text, responseText = '' }) {
+    let title = 'New Chat';
+    this.message_delta = undefined;
+    this.message_start = undefined;
+    const convo = `<initial_message>
+  ${truncateText(text)}
+  </initial_message>
+  <response>
+  ${JSON.stringify(truncateText(responseText))}
+  </response>`;
+
+    const { ANTHROPIC_TITLE_MODEL } = process.env ?? {};
+    const model = this.options.titleModel ?? ANTHROPIC_TITLE_MODEL ?? 'claude-3-haiku-20240307';
+    const system = titleFunctionPrompt;
+
+    const titleChatCompletion = async () => {
+      const content = `<conversation_context>
+  ${convo}
+  </conversation_context>
+  
+  Please generate a title for this conversation.`;
+
+      const titleMessage = { role: 'user', content };
+      const requestOptions = {
+        model,
+        temperature: 0.3,
+        max_tokens: 1024,
+        system,
+        stop_sequences: ['\n\nHuman:', '\n\nAssistant', '</function_calls>'],
+        messages: [titleMessage],
+      };
+
+      try {
+        const response = await this.createResponse(
+          this.getClient(requestOptions),
+          requestOptions,
+          true,
+        );
+        let promptTokens = response?.usage?.input_tokens;
+        let completionTokens = response?.usage?.output_tokens;
+        if (!promptTokens) {
+          promptTokens = this.getTokenCountForMessage(titleMessage);
+          promptTokens += this.getTokenCountForMessage({ role: 'system', content: system });
+        }
+        if (!completionTokens) {
+          completionTokens = this.getTokenCountForMessage(response.content[0]);
+        }
+        await this.recordTokenUsage({
+          model,
+          promptTokens,
+          completionTokens,
+          context: 'title',
+        });
+        const text = response.content[0].text;
+        title = parseParamFromPrompt(text, 'title');
+      } catch (e) {
+        logger.error('[AnthropicClient] There was an issue generating the title', e);
+      }
+    };
+
+    await titleChatCompletion();
+    logger.debug('[AnthropicClient] Convo Title: ' + title);
+    return title;
   }
 }
 

api/app/clients/ChatGPTClient.js

@@ -1,9 +1,20 @@
-const crypto = require('crypto');
 const Keyv = require('keyv');
-const { encoding_for_model: encodingForModel, get_encoding: getEncoding } = require('tiktoken');
+const crypto = require('crypto');
+const { CohereClient } = require('cohere-ai');
 const { fetchEventSource } = require('@waylaidwanderer/fetch-event-source');
-const { Agent, ProxyAgent } = require('undici');
+const { encoding_for_model: encodingForModel, get_encoding: getEncoding } = require('tiktoken');
+const {
+  ImageDetail,
+  EModelEndpoint,
+  resolveHeaders,
+  CohereConstants,
+  mapModelToAzureConfig,
+} = require('librechat-data-provider');
+const { extractBaseURL, constructAzureURL, genAzureChatCompletion } = require('~/utils');
+const { createContextHandlers } = require('./prompts');
+const { createCoherePayload } = require('./llm');
 const BaseClient = require('./BaseClient');
+const { logger } = require('~/config');
 
 const CHATGPT_MODEL = 'gpt-3.5-turbo';
 const tokenizersCache = {};
@@ -140,11 +151,13 @@ class ChatGPTClient extends BaseClient {
     return tokenizer;
   }
 
-  async getCompletion(input, onProgress, abortController = null) {
+  /** @type {getCompletion} */
+  async getCompletion(input, onProgress, onTokenProgress, abortController = null) {
     if (!abortController) {
       abortController = new AbortController();
     }
-    const modelOptions = { ...this.modelOptions };
+
+    let modelOptions = { ...this.modelOptions };
     if (typeof onProgress === 'function') {
       modelOptions.stream = true;
     }
@@ -159,50 +172,178 @@ class ChatGPTClient extends BaseClient {
     }
 
     const { debug } = this.options;
-    const url = this.completionsUrl;
+    let baseURL = this.completionsUrl;
     if (debug) {
       console.debug();
-      console.debug(url);
+      console.debug(baseURL);
       console.debug(modelOptions);
       console.debug();
     }
+
     const opts = {
       method: 'POST',
       headers: {
         'Content-Type': 'application/json',
       },
-      body: JSON.stringify(modelOptions),
-      dispatcher: new Agent({
-        bodyTimeout: 0,
-        headersTimeout: 0,
-      }),
     };
 
-    if (this.apiKey && this.options.azure) {
-      opts.headers['api-key'] = this.apiKey;
+    if (this.isVisionModel) {
+      modelOptions.max_tokens = 4000;
+    }
+
+    /** @type {TAzureConfig | undefined} */
+    const azureConfig = this.options?.req?.app?.locals?.[EModelEndpoint.azureOpenAI];
+
+    const isAzure = this.azure || this.options.azure;
+    if (
+      (isAzure && this.isVisionModel && azureConfig) ||
+      (azureConfig && this.isVisionModel && this.options.endpoint === EModelEndpoint.azureOpenAI)
+    ) {
+      const { modelGroupMap, groupMap } = azureConfig;
+      const {
+        azureOptions,
+        baseURL,
+        headers = {},
+        serverless,
+      } = mapModelToAzureConfig({
+        modelName: modelOptions.model,
+        modelGroupMap,
+        groupMap,
+      });
+      opts.headers = resolveHeaders(headers);
+      this.langchainProxy = extractBaseURL(baseURL);
+      this.apiKey = azureOptions.azureOpenAIApiKey;
+
+      const groupName = modelGroupMap[modelOptions.model].group;
+      this.options.addParams = azureConfig.groupMap[groupName].addParams;
+      this.options.dropParams = azureConfig.groupMap[groupName].dropParams;
+      // Note: `forcePrompt` not re-assigned as only chat models are vision models
+
+      this.azure = !serverless && azureOptions;
+      this.azureEndpoint =
+        !serverless && genAzureChatCompletion(this.azure, modelOptions.model, this);
+      if (serverless === true) {
+        this.options.defaultQuery = azureOptions.azureOpenAIApiVersion
+          ? { 'api-version': azureOptions.azureOpenAIApiVersion }
+          : undefined;
+        this.options.headers['api-key'] = this.apiKey;
+      }
+    }
+
+    if (this.options.defaultQuery) {
+      opts.defaultQuery = this.options.defaultQuery;
+    }
+
+    if (this.options.headers) {
+      opts.headers = { ...opts.headers, ...this.options.headers };
+    }
+
+    if (isAzure) {
+      // Azure does not accept `model` in the body, so we need to remove it.
+      delete modelOptions.model;
+
+      baseURL = this.langchainProxy
+        ? constructAzureURL({
+          baseURL: this.langchainProxy,
+          azureOptions: this.azure,
+        })
+        : this.azureEndpoint.split(/(?<!\/)\/(chat|completion)\//)[0];
+
+      if (this.options.forcePrompt) {
+        baseURL += '/completions';
+      } else {
+        baseURL += '/chat/completions';
+      }
+
+      opts.defaultQuery = { 'api-version': this.azure.azureOpenAIApiVersion };
+      opts.headers = { ...opts.headers, 'api-key': this.apiKey };
     } else if (this.apiKey) {
       opts.headers.Authorization = `Bearer ${this.apiKey}`;
     }
 
+    if (process.env.OPENAI_ORGANIZATION) {
+      opts.headers['OpenAI-Organization'] = process.env.OPENAI_ORGANIZATION;
+    }
+
     if (this.useOpenRouter) {
       opts.headers['HTTP-Referer'] = 'https://librechat.ai';
       opts.headers['X-Title'] = 'LibreChat';
     }
 
-    if (this.options.headers) {
-      opts.headers = { ...opts.headers, ...this.options.headers };
+    /* hacky fixes for Mistral AI API:
+      - Re-orders system message to the top of the messages payload, as not allowed anywhere else
+      - If there is only one message and it's a system message, change the role to user
+      */
+    if (baseURL.includes('https://api.mistral.ai/v1') && modelOptions.messages) {
+      const { messages } = modelOptions;
+
+      const systemMessageIndex = messages.findIndex((msg) => msg.role === 'system');
+
+      if (systemMessageIndex > 0) {
+        const [systemMessage] = messages.splice(systemMessageIndex, 1);
+        messages.unshift(systemMessage);
+      }
+
+      modelOptions.messages = messages;
+
+      if (messages.length === 1 && messages[0].role === 'system') {
+        modelOptions.messages[0].role = 'user';
+      }
     }
 
-    if (this.options.proxy) {
-      opts.dispatcher = new ProxyAgent(this.options.proxy);
+    if (this.options.addParams && typeof this.options.addParams === 'object') {
+      modelOptions = {
+        ...modelOptions,
+        ...this.options.addParams,
+      };
+      logger.debug('[ChatGPTClient] chatCompletion: added params', {
+        addParams: this.options.addParams,
+        modelOptions,
+      });
     }
 
+    if (this.options.dropParams && Array.isArray(this.options.dropParams)) {
+      this.options.dropParams.forEach((param) => {
+        delete modelOptions[param];
+      });
+      logger.debug('[ChatGPTClient] chatCompletion: dropped params', {
+        dropParams: this.options.dropParams,
+        modelOptions,
+      });
+    }
+
+    if (baseURL.startsWith(CohereConstants.API_URL)) {
+      const payload = createCoherePayload({ modelOptions });
+      return await this.cohereChatCompletion({ payload, onTokenProgress });
+    }
+
+    if (baseURL.includes('v1') && !baseURL.includes('/completions') && !this.isChatCompletion) {
+      baseURL = baseURL.split('v1')[0] + 'v1/completions';
+    } else if (
+      baseURL.includes('v1') &&
+      !baseURL.includes('/chat/completions') &&
+      this.isChatCompletion
+    ) {
+      baseURL = baseURL.split('v1')[0] + 'v1/chat/completions';
+    }
+
+    const BASE_URL = new URL(baseURL);
+    if (opts.defaultQuery) {
+      Object.entries(opts.defaultQuery).forEach(([key, value]) => {
+        BASE_URL.searchParams.append(key, value);
+      });
+      delete opts.defaultQuery;
+    }
+
+    const completionsURL = BASE_URL.toString();
+    opts.body = JSON.stringify(modelOptions);
+
     if (modelOptions.stream) {
       // eslint-disable-next-line no-async-promise-executor
       return new Promise(async (resolve, reject) => {
         try {
           let done = false;
-          await fetchEventSource(url, {
+          await fetchEventSource(completionsURL, {
             ...opts,
             signal: abortController.signal,
             async onopen(response) {
@@ -230,7 +371,6 @@ class ChatGPTClient extends BaseClient {
               // workaround for private API not sending [DONE] event
               if (!done) {
                 onProgress('[DONE]');
-                abortController.abort();
                 resolve();
               }
             },
@@ -243,14 +383,13 @@ class ChatGPTClient extends BaseClient {
             },
             onmessage(message) {
               if (debug) {
-                // console.debug(message);
+                console.debug(message);
               }
               if (!message.data || message.event === 'ping') {
                 return;
               }
               if (message.data === '[DONE]') {
                 onProgress('[DONE]');
-                abortController.abort();
                 resolve();
                 done = true;
                 return;
@@ -263,7 +402,7 @@ class ChatGPTClient extends BaseClient {
         }
       });
     }
-    const response = await fetch(url, {
+    const response = await fetch(completionsURL, {
       ...opts,
       signal: abortController.signal,
     });
@@ -281,6 +420,43 @@ class ChatGPTClient extends BaseClient {
     return response.json();
   }
 
+  /** @type {cohereChatCompletion} */
+  async cohereChatCompletion({ payload, onTokenProgress }) {
+    const cohere = new CohereClient({
+      token: this.apiKey,
+      environment: this.completionsUrl,
+    });
+
+    if (!payload.stream) {
+      const chatResponse = await cohere.chat(payload);
+      return chatResponse.text;
+    }
+
+    const chatStream = await cohere.chatStream(payload);
+    let reply = '';
+    for await (const message of chatStream) {
+      if (!message) {
+        continue;
+      }
+
+      if (message.eventType === 'text-generation' && message.text) {
+        onTokenProgress(message.text);
+        reply += message.text;
+      }
+      /*
+      Cohere API Chinese Unicode character replacement hotfix.
+      Should be un-commented when the following issue is resolved:
+      https://github.com/cohere-ai/cohere-typescript/issues/151
+
+      else if (message.eventType === 'stream-end' && message.response) {
+        reply = message.response.text;
+      }
+      */
+    }
+
+    return reply;
+  }
+
   async generateTitle(userMessage, botMessage) {
     const instructionsPayload = {
       role: 'system',
@@ -439,26 +615,70 @@ ${botMessage.message}
 
   async buildPrompt(messages, { isChatGptModel = false, promptPrefix = null }) {
     promptPrefix = (promptPrefix || this.options.promptPrefix || '').trim();
+
+    // Handle attachments and create augmentedPrompt
+    if (this.options.attachments) {
+      const attachments = await this.options.attachments;
+      const lastMessage = messages[messages.length - 1];
+
+      if (this.message_file_map) {
+        this.message_file_map[lastMessage.messageId] = attachments;
+      } else {
+        this.message_file_map = {
+          [lastMessage.messageId]: attachments,
+        };
+      }
+
+      const files = await this.addImageURLs(lastMessage, attachments);
+      this.options.attachments = files;
+
+      this.contextHandlers = createContextHandlers(this.options.req, lastMessage.text);
+    }
+
+    if (this.message_file_map) {
+      this.contextHandlers = createContextHandlers(
+        this.options.req,
+        messages[messages.length - 1].text,
+      );
+    }
+
+    // Calculate image token cost and process embedded files
+    messages.forEach((message, i) => {
+      if (this.message_file_map && this.message_file_map[message.messageId]) {
+        const attachments = this.message_file_map[message.messageId];
+        for (const file of attachments) {
+          if (file.embedded) {
+            this.contextHandlers?.processFile(file);
+            continue;
+          }
+
+          messages[i].tokenCount =
+            (messages[i].tokenCount || 0) +
+            this.calculateImageTokenCost({
+              width: file.width,
+              height: file.height,
+              detail: this.options.imageDetail ?? ImageDetail.auto,
+            });
+        }
+      }
+    });
+
+    if (this.contextHandlers) {
+      this.augmentedPrompt = await this.contextHandlers.createContext();
+      promptPrefix = this.augmentedPrompt + promptPrefix;
+    }
+
     if (promptPrefix) {
       // If the prompt prefix doesn't end with the end token, add it.
       if (!promptPrefix.endsWith(`${this.endToken}`)) {
         promptPrefix = `${promptPrefix.trim()}${this.endToken}\n\n`;
       }
       promptPrefix = `${this.startToken}Instructions:\n${promptPrefix}`;
-    } else {
-      const currentDateString = new Date().toLocaleDateString('en-us', {
-        year: 'numeric',
-        month: 'long',
-        day: 'numeric',
-      });
-      promptPrefix = `${this.startToken}Instructions:\nYou are ChatGPT, a large language model trained by OpenAI. Respond conversationally.\nCurrent date: ${currentDateString}${this.endToken}\n\n`;
     }
-
     const promptSuffix = `${this.startToken}${this.chatGptLabel}:\n`; // Prompt ChatGPT to respond.
 
     const instructionsPayload = {
       role: 'system',
-      name: 'instructions',
       content: promptPrefix,
     };
 
@@ -541,14 +761,10 @@ ${botMessage.message}
       this.maxResponseTokens,
     );
 
-    if (this.options.debug) {
-      console.debug(`Prompt : ${prompt}`);
-    }
-
     if (isChatGptModel) {
       return { prompt: [instructionsPayload, messagePayload], context };
     }
-    return { prompt, context };
+    return { prompt, context, promptTokens: currentTokenCount };
   }
 
   getTokenCount(text) {

api/app/clients/OllamaClient.js

@@ -0,0 +1,161 @@
+const { z } = require('zod');
+const axios = require('axios');
+const { Ollama } = require('ollama');
+const { Constants } = require('librechat-data-provider');
+const { deriveBaseURL, logAxiosError } = require('~/utils');
+const { sleep } = require('~/server/utils');
+const { logger } = require('~/config');
+
+const ollamaPayloadSchema = z.object({
+  mirostat: z.number().optional(),
+  mirostat_eta: z.number().optional(),
+  mirostat_tau: z.number().optional(),
+  num_ctx: z.number().optional(),
+  repeat_last_n: z.number().optional(),
+  repeat_penalty: z.number().optional(),
+  temperature: z.number().optional(),
+  seed: z.number().nullable().optional(),
+  stop: z.array(z.string()).optional(),
+  tfs_z: z.number().optional(),
+  num_predict: z.number().optional(),
+  top_k: z.number().optional(),
+  top_p: z.number().optional(),
+  stream: z.optional(z.boolean()),
+  model: z.string(),
+});
+
+/**
+ * @param {string} imageUrl
+ * @returns {string}
+ * @throws {Error}
+ */
+const getValidBase64 = (imageUrl) => {
+  const parts = imageUrl.split(';base64,');
+
+  if (parts.length === 2) {
+    return parts[1];
+  } else {
+    logger.error('Invalid or no Base64 string found in URL.');
+  }
+};
+
+class OllamaClient {
+  constructor(options = {}) {
+    const host = deriveBaseURL(options.baseURL ?? 'http://localhost:11434');
+    this.streamRate = options.streamRate ?? Constants.DEFAULT_STREAM_RATE;
+    /** @type {Ollama} */
+    this.client = new Ollama({ host });
+  }
+
+  /**
+   * Fetches Ollama models from the specified base API path.
+   * @param {string} baseURL
+   * @returns {Promise<string[]>} The Ollama models.
+   */
+  static async fetchModels(baseURL) {
+    let models = [];
+    if (!baseURL) {
+      return models;
+    }
+    try {
+      const ollamaEndpoint = deriveBaseURL(baseURL);
+      /** @type {Promise<AxiosResponse<OllamaListResponse>>} */
+      const response = await axios.get(`${ollamaEndpoint}/api/tags`, {
+        timeout: 5000,
+      });
+      models = response.data.models.map((tag) => tag.name);
+      return models;
+    } catch (error) {
+      const logMessage =
+        'Failed to fetch models from Ollama API. If you are not using Ollama directly, and instead, through some aggregator or reverse proxy that handles fetching via OpenAI spec, ensure the name of the endpoint doesn\'t start with `ollama` (case-insensitive).';
+      logAxiosError({ message: logMessage, error });
+      return [];
+    }
+  }
+
+  /**
+   * @param {ChatCompletionMessage[]} messages
+   * @returns {OllamaMessage[]}
+   */
+  static formatOpenAIMessages(messages) {
+    const ollamaMessages = [];
+
+    for (const message of messages) {
+      if (typeof message.content === 'string') {
+        ollamaMessages.push({
+          role: message.role,
+          content: message.content,
+        });
+        continue;
+      }
+
+      let aggregatedText = '';
+      let imageUrls = [];
+
+      for (const content of message.content) {
+        if (content.type === 'text') {
+          aggregatedText += content.text + ' ';
+        } else if (content.type === 'image_url') {
+          imageUrls.push(getValidBase64(content.image_url.url));
+        }
+      }
+
+      const ollamaMessage = {
+        role: message.role,
+        content: aggregatedText.trim(),
+      };
+
+      if (imageUrls.length > 0) {
+        ollamaMessage.images = imageUrls;
+      }
+
+      ollamaMessages.push(ollamaMessage);
+    }
+
+    return ollamaMessages;
+  }
+
+  /***
+   * @param {Object} params
+   * @param {ChatCompletionPayload} params.payload
+   * @param {onTokenProgress} params.onProgress
+   * @param {AbortController} params.abortController
+   */
+  async chatCompletion({ payload, onProgress, abortController = null }) {
+    let intermediateReply = '';
+
+    const parameters = ollamaPayloadSchema.parse(payload);
+    const messages = OllamaClient.formatOpenAIMessages(payload.messages);
+
+    if (parameters.stream) {
+      const stream = await this.client.chat({
+        messages,
+        ...parameters,
+      });
+
+      for await (const chunk of stream) {
+        const token = chunk.message.content;
+        intermediateReply += token;
+        onProgress(token);
+        if (abortController.signal.aborted) {
+          stream.controller.abort();
+          break;
+        }
+
+        await sleep(this.streamRate);
+      }
+    }
+    // TODO: regular completion
+    else {
+      // const generation = await this.client.generate(payload);
+    }
+
+    return intermediateReply;
+  }
+  catch(err) {
+    logger.error('[OllamaClient.chatCompletion]', err);
+    throw err;
+  }
+}
+
+module.exports = { OllamaClient, ollamaPayloadSchema };

api/app/clients/PluginsClient.js

@@ -1,14 +1,16 @@
 const OpenAIClient = require('./OpenAIClient');
-const { CallbackManager } = require('langchain/callbacks');
+const { CallbackManager } = require('@langchain/core/callbacks/manager');
 const { BufferMemory, ChatMessageHistory } = require('langchain/memory');
-const { initializeCustomAgent, initializeFunctionsAgent } = require('./agents');
 const { addImages, buildErrorInput, buildPromptPrefix } = require('./output_parsers');
-const checkBalance = require('../../models/checkBalance');
+const { initializeCustomAgent, initializeFunctionsAgent } = require('./agents');
+const { processFileURL } = require('~/server/services/Files/process');
+const { EModelEndpoint } = require('librechat-data-provider');
 const { formatLangChainMessages } = require('./prompts');
-const { isEnabled } = require('../../server/utils');
-const { extractBaseURL } = require('../../utils');
-const { SelfReflectionTool } = require('./tools');
+const checkBalance = require('~/models/checkBalance');
+const { isEnabled } = require('~/server/utils');
+const { extractBaseURL } = require('~/utils');
 const { loadTools } = require('./tools/util');
+const { logger } = require('~/config');
 
 class PluginsClient extends OpenAIClient {
   constructor(apiKey, options = {}) {
@@ -28,27 +30,25 @@ class PluginsClient extends OpenAIClient {
 
     super.setOptions(options);
 
-    if (this.functionsAgent && this.agentOptions.model && !this.useOpenRouter) {
-      this.agentOptions.model = this.getFunctionModelName(this.agentOptions.model);
-    }
-
     this.isGpt3 = this.modelOptions?.model?.includes('gpt-3');
 
     if (this.options.reverseProxyUrl) {
       this.langchainProxy = extractBaseURL(this.options.reverseProxyUrl);
-      !this.langchainProxy &&
-        console.warn(`The reverse proxy URL ${this.options.reverseProxyUrl} is not valid for Plugins.
-The url must follow OpenAI specs, for example: https://localhost:8080/v1/chat/completions
-If your reverse proxy is compatible to OpenAI specs in every other way, it may still work without plugins enabled.`);
     }
   }
 
   getSaveOptions() {
     return {
+      artifacts: this.options.artifacts,
       chatGptLabel: this.options.chatGptLabel,
+      modelLabel: this.options.modelLabel,
       promptPrefix: this.options.promptPrefix,
+      tools: this.options.tools,
       ...this.modelOptions,
       agentOptions: this.agentOptions,
+      iconURL: this.options.iconURL,
+      greeting: this.options.greeting,
+      spec: this.options.spec,
     };
   }
 
@@ -57,7 +57,9 @@ If your reverse proxy is compatible to OpenAI specs in every other way, it may s
   }
 
   getFunctionModelName(input) {
-    if (input.includes('gpt-3.5-turbo')) {
+    if (/-(?!0314)\d{4}/.test(input)) {
+      return input;
+    } else if (input.includes('gpt-3.5-turbo')) {
       return 'gpt-3.5-turbo';
     } else if (input.includes('gpt-4')) {
       return 'gpt-4';
@@ -86,17 +88,15 @@ If your reverse proxy is compatible to OpenAI specs in every other way, it may s
       initialMessageCount: this.currentMessages.length + 1,
     });
 
-    if (this.options.debug) {
-      console.debug(
-        `<-----Agent Model: ${model.modelName} | Temp: ${model.temperature} | Functions: ${this.functionsAgent}----->`,
-      );
-    }
+    logger.debug(
+      `[PluginsClient] Agent Model: ${model.modelName} | Temp: ${model.temperature} | Functions: ${this.functionsAgent}`,
+    );
 
     // Map Messages to Langchain format
     const pastMessages = formatLangChainMessages(this.currentMessages.slice(0, -1), {
       userName: this.options?.name,
     });
-    this.options.debug && console.debug('pastMessages: ', pastMessages);
+    logger.debug('[PluginsClient] pastMessages: ' + pastMessages.length);
 
     // TODO: use readOnly memory, TokenBufferMemory? (both unavailable in LangChainJS)
     const memory = new BufferMemory({
@@ -104,7 +104,7 @@ If your reverse proxy is compatible to OpenAI specs in every other way, it may s
       chatHistory: new ChatMessageHistory(pastMessages),
     });
 
-    this.tools = await loadTools({
+    const { loadedTools } = await loadTools({
       user,
       model,
       tools: this.options.tools,
@@ -114,30 +114,29 @@ If your reverse proxy is compatible to OpenAI specs in every other way, it may s
         signal: this.abortController.signal,
         openAIApiKey: this.openAIApiKey,
         conversationId: this.conversationId,
-        debug: this.options?.debug,
+        fileStrategy: this.options.req.app.locals.fileStrategy,
+        processFileURL,
         message,
       },
+      useSpecs: true,
     });
 
-    if (this.tools.length > 0 && !this.functionsAgent) {
-      this.tools.push(new SelfReflectionTool({ message, isGpt3: false }));
-    } else if (this.tools.length === 0) {
+    if (loadedTools.length === 0) {
       return;
     }
 
-    if (this.options.debug) {
-      console.debug('Requested Tools');
-      console.debug(this.options.tools);
-      console.debug('Loaded Tools');
-      console.debug(this.tools.map((tool) => tool.name));
-    }
+    this.tools = loadedTools;
+
+    logger.debug('[PluginsClient] Requested Tools', this.options.tools);
+    logger.debug(
+      '[PluginsClient] Loaded Tools',
+      this.tools.map((tool) => tool.name),
+    );
 
     const handleAction = (action, runId, callback = null) => {
       this.saveLatestAction(action);
 
-      if (this.options.debug) {
-        console.debug('Latest Agent Action ', this.actions[this.actions.length - 1]);
-      }
+      logger.debug('[PluginsClient] Latest Agent Action ', this.actions[this.actions.length - 1]);
 
       if (typeof callback === 'function') {
         callback(action, runId);
@@ -146,14 +145,22 @@ If your reverse proxy is compatible to OpenAI specs in every other way, it may s
 
     // initialize agent
     const initializer = this.functionsAgent ? initializeFunctionsAgent : initializeCustomAgent;
+
+    let customInstructions = (this.options.promptPrefix ?? '').trim();
+    if (typeof this.options.artifactsPrompt === 'string' && this.options.artifactsPrompt) {
+      customInstructions = `${customInstructions ?? ''}\n${this.options.artifactsPrompt}`.trim();
+    }
+
     this.executor = await initializer({
       model,
       signal,
       pastMessages,
       tools: this.tools,
-      currentDateString: this.currentDateString,
+      customInstructions,
       verbose: this.options.debug,
       returnIntermediateSteps: true,
+      customName: this.options.chatGptLabel,
+      currentDateString: this.currentDateString,
       callbackManager: CallbackManager.fromHandlers({
         async handleAgentAction(action, runId) {
           handleAction(action, runId, onAgentAction);
@@ -166,9 +173,7 @@ If your reverse proxy is compatible to OpenAI specs in every other way, it may s
       }),
     });
 
-    if (this.options.debug) {
-      console.debug('Loaded agent.');
-    }
+    logger.debug('[PluginsClient] Loaded agent.');
   }
 
   async executorCall(message, { signal, stream, onToolStart, onToolEnd }) {
@@ -184,12 +189,10 @@ If your reverse proxy is compatible to OpenAI specs in every other way, it may s
       });
       const input = attempts > 1 ? errorInput : message;
 
-      if (this.options.debug) {
-        console.debug(`Attempt ${attempts} of ${maxAttempts}`);
-      }
+      logger.debug(`[PluginsClient] Attempt ${attempts} of ${maxAttempts}`);
 
-      if (this.options.debug && errorMessage.length > 0) {
-        console.debug('Caught error, input:', input);
+      if (errorMessage.length > 0) {
+        logger.debug('[PluginsClient] Caught error, input: ' + JSON.stringify(input));
       }
 
       try {
@@ -212,10 +215,10 @@ If your reverse proxy is compatible to OpenAI specs in every other way, it may s
         ]);
         break; // Exit the loop if the function call is successful
       } catch (err) {
-        console.error(err);
+        logger.error('[PluginsClient] executorCall error:', err);
         if (attempts === maxAttempts) {
           const { run } = this.runManager.getRunByConversationId(this.conversationId);
-          const defaultOutput = `Encountered an error while attempting to respond. Error: ${err.message}`;
+          const defaultOutput = `Encountered an error while attempting to respond: ${err.message}`;
           this.result.output = run && run.error ? run.error : defaultOutput;
           this.result.errorMessage = run && run.error ? run.error : err.message;
           this.result.intermediateSteps = this.actions;
@@ -225,10 +228,20 @@ If your reverse proxy is compatible to OpenAI specs in every other way, it may s
     }
   }
 
+  /**
+   *
+   * @param {TMessage} responseMessage
+   * @param {Partial<TMessage>} saveOptions
+   * @param {string} user
+   * @returns
+   */
   async handleResponseMessage(responseMessage, saveOptions, user) {
     const { output, errorMessage, ...result } = this.result;
-    this.options.debug &&
-      console.debug('[handleResponseMessage] Output:', { output, errorMessage, ...result });
+    logger.debug('[PluginsClient][handleResponseMessage] Output:', {
+      output,
+      errorMessage,
+      ...result,
+    });
     const { error } = responseMessage;
     if (!error) {
       responseMessage.tokenCount = this.getTokenCountForResponse(responseMessage);
@@ -240,22 +253,33 @@ If your reverse proxy is compatible to OpenAI specs in every other way, it may s
       await this.recordTokenUsage(responseMessage);
     }
 
-    await this.saveMessageToDatabase(responseMessage, saveOptions, user);
+    this.responsePromise = this.saveMessageToDatabase(responseMessage, saveOptions, user);
     delete responseMessage.tokenCount;
     return { ...responseMessage, ...result };
   }
 
   async sendMessage(message, opts = {}) {
+    /** @type {{ filteredTools: string[], includedTools: string[] }} */
+    const { filteredTools = [], includedTools = [] } = this.options.req.app.locals;
+
+    if (includedTools.length > 0) {
+      const tools = this.options.tools.filter((plugin) => includedTools.includes(plugin));
+      this.options.tools = tools;
+    } else {
+      const tools = this.options.tools.filter((plugin) => !filteredTools.includes(plugin));
+      this.options.tools = tools;
+    }
+
     // If a message is edited, no tools can be used.
     const completionMode = this.options.tools.length === 0 || opts.isEdited;
     if (completionMode) {
       this.setOptions(opts);
       return super.sendMessage(message, opts);
     }
-    this.options.debug && console.log('Plugins sendMessage', message, opts);
+
+    logger.debug('[PluginsClient] sendMessage', { userMessageText: message, opts });
     const {
       user,
-      isEdited,
       conversationId,
       responseMessageId,
       saveOptions,
@@ -266,6 +290,14 @@ If your reverse proxy is compatible to OpenAI specs in every other way, it may s
       onToolEnd,
     } = await this.handleStartMethods(message, opts);
 
+    if (opts.progressCallback) {
+      opts.onProgress = opts.progressCallback.call(null, {
+        ...(opts.progressOptions ?? {}),
+        parentMessageId: userMessage.messageId,
+        messageId: responseMessageId,
+      });
+    }
+
     this.currentMessages.push(userMessage);
 
     let {
@@ -282,10 +314,10 @@ If your reverse proxy is compatible to OpenAI specs in every other way, it may s
     );
 
     if (tokenCountMap) {
-      console.dir(tokenCountMap, { depth: null });
+      logger.debug('[PluginsClient] tokenCountMap', { tokenCountMap });
       if (tokenCountMap[userMessage.messageId]) {
         userMessage.tokenCount = tokenCountMap[userMessage.messageId];
-        console.log('userMessage.tokenCount', userMessage.tokenCount);
+        logger.debug('[PluginsClient] userMessage.tokenCount', userMessage.tokenCount);
       }
       this.handleTokenCountMap(tokenCountMap);
     }
@@ -294,7 +326,15 @@ If your reverse proxy is compatible to OpenAI specs in every other way, it may s
     if (payload) {
       this.currentMessages = payload;
     }
-    await this.saveMessageToDatabase(userMessage, saveOptions, user);
+
+    if (!this.skipSaveUserMessage) {
+      this.userMessagePromise = this.saveMessageToDatabase(userMessage, saveOptions, user);
+      if (typeof opts?.getReqData === 'function') {
+        opts.getReqData({
+          userMessagePromise: this.userMessagePromise,
+        });
+      }
+    }
 
     if (isEnabled(process.env.CHECK_BALANCE)) {
       await checkBalance({
@@ -306,16 +346,18 @@ If your reverse proxy is compatible to OpenAI specs in every other way, it may s
           amount: promptTokens,
           debug: this.options.debug,
           model: this.modelOptions.model,
+          endpoint: EModelEndpoint.openAI,
         },
       });
     }
 
     const responseMessage = {
+      endpoint: EModelEndpoint.gptPlugins,
+      iconURL: this.options.iconURL,
       messageId: responseMessageId,
       conversationId,
       parentMessageId: userMessage.messageId,
       isCreatedByUser: false,
-      isEdited,
       model: this.modelOptions.model,
       sender: this.sender,
       promptTokens,
@@ -370,10 +412,7 @@ If your reverse proxy is compatible to OpenAI specs in every other way, it may s
       return await this.handleResponseMessage(responseMessage, saveOptions, user);
     }
 
-    if (this.options.debug) {
-      console.debug('Plugins completion phase: this.result');
-      console.debug(this.result);
-    }
+    logger.debug('[PluginsClient] Completion phase: this.result', this.result);
 
     const promptPrefix = buildPromptPrefix({
       result: this.result,
@@ -381,28 +420,20 @@ If your reverse proxy is compatible to OpenAI specs in every other way, it may s
       functionsAgent: this.functionsAgent,
     });
 
-    if (this.options.debug) {
-      console.debug('Plugins: promptPrefix');
-      console.debug(promptPrefix);
-    }
+    logger.debug('[PluginsClient]', { promptPrefix });
 
     payload = await this.buildCompletionPrompt({
       messages: this.currentMessages,
       promptPrefix,
     });
 
-    if (this.options.debug) {
-      console.debug('buildCompletionPrompt Payload');
-      console.debug(payload);
-    }
+    logger.debug('[PluginsClient] buildCompletionPrompt Payload', payload);
     responseMessage.text = await this.sendCompletion(payload, opts);
     return await this.handleResponseMessage(responseMessage, saveOptions, user);
   }
 
   async buildCompletionPrompt({ messages, promptPrefix: _promptPrefix }) {
-    if (this.options.debug) {
-      console.debug('buildCompletionPrompt messages', messages);
-    }
+    logger.debug('[PluginsClient] buildCompletionPrompt messages', messages);
 
     const orderedMessages = messages;
     let promptPrefix = _promptPrefix.trim();
@@ -415,7 +446,6 @@ If your reverse proxy is compatible to OpenAI specs in every other way, it may s
 
     const instructionsPayload = {
       role: 'system',
-      name: 'instructions',
       content: promptPrefix,
     };
 

api/app/clients/TextStream.js

@@ -1,4 +1,5 @@
 const { Readable } = require('stream');
+const { logger } = require('~/config');
 
 class TextStream extends Readable {
   constructor(text, options = {}) {
@@ -38,7 +39,7 @@ class TextStream extends Readable {
       });
 
       this.on('end', () => {
-        // console.log('Stream ended');
+        // logger.debug('[processTextStream] Stream ended');
         resolve();
       });
 
@@ -50,7 +51,7 @@ class TextStream extends Readable {
     try {
       await streamPromise;
     } catch (err) {
-      console.error('Error processing text stream:', err);
+      logger.error('[processTextStream] Error in text stream:', err);
       // Handle the error appropriately, e.g., return an error message or throw an error
     }
   }

api/app/clients/agents/CustomAgent/CustomAgent.js

@@ -1,5 +1,5 @@
 const { ZeroShotAgent } = require('langchain/agents');
-const { PromptTemplate, renderTemplate } = require('langchain/prompts');
+const { PromptTemplate, renderTemplate } = require('@langchain/core/prompts');
 const { gpt3, gpt4 } = require('./instructions');
 
 class CustomAgent extends ZeroShotAgent {

api/app/clients/agents/CustomAgent/initializeCustomAgent.js

@@ -7,16 +7,24 @@ const {
   ChatPromptTemplate,
   SystemMessagePromptTemplate,
   HumanMessagePromptTemplate,
-} = require('langchain/prompts');
+} = require('@langchain/core/prompts');
 
 const initializeCustomAgent = async ({
   tools,
   model,
   pastMessages,
+  customName,
+  customInstructions,
   currentDateString,
   ...rest
 }) => {
   let prompt = CustomAgent.createPrompt(tools, { currentDateString, model: model.modelName });
+  if (customName) {
+    prompt = `You are "${customName}".\n${prompt}`;
+  }
+  if (customInstructions) {
+    prompt = `${prompt}\n${customInstructions}`;
+  }
 
   const chatPrompt = ChatPromptTemplate.fromMessages([
     new SystemMessagePromptTemplate(prompt),

api/app/clients/agents/CustomAgent/instructions.js

@@ -1,44 +1,3 @@
-/*
-module.exports = `You are ChatGPT, a Large Language model with useful tools.
-
-Talk to the human and provide meaningful answers when questions are asked.
-
-Use the tools when you need them, but use your own knowledge if you are confident of the answer. Keep answers short and concise.
-
-A tool is not usually needed for creative requests, so do your best to answer them without tools.
-
-Avoid repeating identical answers if it appears before. Only fulfill the human's requests, do not create extra steps beyond what the human has asked for.
-
-Your input for 'Action' should be the name of tool used only.
-
-Be honest. If you can't answer something, or a tool is not appropriate, say you don't know or answer to the best of your ability.
-
-Attempt to fulfill the human's requests in as few actions as possible`;
-*/
-
-// module.exports = `You are ChatGPT, a highly knowledgeable and versatile large language model.
-
-// Engage with the Human conversationally, providing concise and meaningful answers to questions. Utilize built-in tools when necessary, except for creative requests, where relying on your own knowledge is preferred. Aim for variety and avoid repetitive answers.
-
-// For your 'Action' input, state the name of the tool used only, and honor user requests without adding extra steps. Always be honest; if you cannot provide an appropriate answer or tool, admit that or do your best.
-
-// Strive to meet the user's needs efficiently with minimal actions.`;
-
-// import {
-//   BasePromptTemplate,
-//   BaseStringPromptTemplate,
-//   SerializedBasePromptTemplate,
-//   renderTemplate,
-// } from "langchain/prompts";
-
-// prefix: `You are ChatGPT, a highly knowledgeable and versatile large language model.
-// Your objective is to help users by understanding their intent and choosing the best action. Prioritize direct, specific responses. Use concise, varied answers and rely on your knowledge for creative tasks. Utilize tools when needed, and structure results for machine compatibility.
-// prefix: `Objective: to comprehend human intentions based on user input and available tools. Goal: identify the best action to directly address the human's query. In your subsequent steps, you will utilize the chosen action. You may select multiple actions and list them in a meaningful order. Prioritize actions that directly relate to the user's query over general ones. Ensure that the generated thought is highly specific and explicit to best match the user's expectations. Construct the result in a manner that an online open-API would most likely expect. Provide concise and meaningful answers to human queries. Utilize tools when necessary. Relying on your own knowledge is preferred for creative requests. Aim for variety and avoid repetitive answers.
-
-// # Available Actions & Tools:
-// N/A: no suitable action, use your own knowledge.`,
-// suffix: `Remember, all your responses MUST adhere to the described format and only respond if the format is followed. Output exactly with the requested format, avoiding any other text as this will be parsed by a machine. Following 'Action:', provide only one of the actions listed above. If a tool is not necessary, deduce this quickly and finish your response. Honor the human's requests without adding extra steps. Carry out tasks in the sequence written by the human. Always be honest; if you cannot provide an appropriate answer or tool, do your best with your own knowledge. Strive to meet the user's needs efficiently with minimal actions.`;
-
 module.exports = {
   'gpt3-v1': {
     prefix: `Objective: Understand human intentions using user input and available tools. Goal: Identify the most suitable actions to directly address user queries.

api/app/clients/agents/CustomAgent/outputParser.js

@@ -1,4 +1,5 @@
 const { ZeroShotAgentOutputParser } = require('langchain/agents');
+const { logger } = require('~/config');
 
 class CustomOutputParser extends ZeroShotAgentOutputParser {
   constructor(fields) {
@@ -64,9 +65,9 @@ class CustomOutputParser extends ZeroShotAgentOutputParser {
     const match = this.actionValues.exec(text); // old v2
 
     if (!match) {
-      console.log(
-        '\n\n<----------------------HIT NO MATCH PARSING ERROR---------------------->\n\n',
-        match,
+      logger.debug(
+        '\n\n<----------------------[CustomOutputParser] HIT NO MATCH PARSING ERROR---------------------->\n\n' +
+          match,
       );
       const thoughts = text.replace(/[tT]hought:/, '').split('\n');
       // return {
@@ -84,9 +85,9 @@ class CustomOutputParser extends ZeroShotAgentOutputParser {
     let selectedTool = match?.[1].trim().toLowerCase();
 
     if (match && selectedTool === 'n/a') {
-      console.log(
-        '\n\n<----------------------HIT N/A PARSING ERROR---------------------->\n\n',
-        match,
+      logger.debug(
+        '\n\n<----------------------[CustomOutputParser] HIT N/A PARSING ERROR---------------------->\n\n' +
+          match,
       );
       return {
         tool: 'self-reflection',
@@ -97,25 +98,25 @@ class CustomOutputParser extends ZeroShotAgentOutputParser {
 
     let toolIsValid = this.checkIfValidTool(selectedTool);
     if (match && !toolIsValid) {
-      console.log(
-        '\n\n<----------------Tool invalid: Re-assigning Selected Tool---------------->\n\n',
-        match,
+      logger.debug(
+        '\n\n<----------------[CustomOutputParser] Tool invalid: Re-assigning Selected Tool---------------->\n\n' +
+          match,
       );
       selectedTool = this.getValidTool(selectedTool);
     }
 
     if (match && !selectedTool) {
-      console.log(
-        '\n\n<----------------------HIT INVALID TOOL PARSING ERROR---------------------->\n\n',
-        match,
+      logger.debug(
+        '\n\n<----------------------[CustomOutputParser] HIT INVALID TOOL PARSING ERROR---------------------->\n\n' +
+          match,
       );
       selectedTool = 'self-reflection';
     }
 
     if (match && !match[2]) {
-      console.log(
-        '\n\n<----------------------HIT NO ACTION INPUT PARSING ERROR---------------------->\n\n',
-        match,
+      logger.debug(
+        '\n\n<----------------------[CustomOutputParser] HIT NO ACTION INPUT PARSING ERROR---------------------->\n\n' +
+          match,
       );
 
       // In case there is no action input, let's double-check if there is an action input in 'text' variable
@@ -139,7 +140,9 @@ class CustomOutputParser extends ZeroShotAgentOutputParser {
     }
 
     if (match && selectedTool.length > this.longestToolName.length) {
-      console.log('\n\n<----------------------HIT LONG PARSING ERROR---------------------->\n\n');
+      logger.debug(
+        '\n\n<----------------------[CustomOutputParser] HIT LONG PARSING ERROR---------------------->\n\n',
+      );
 
       let action, input, thought;
       let firstIndex = Infinity;
@@ -156,9 +159,9 @@ class CustomOutputParser extends ZeroShotAgentOutputParser {
       // In case there is no action input, let's double-check if there is an action input in 'text' variable
       const actionInputMatch = this.actionInputRegex.exec(text);
       if (action && actionInputMatch) {
-        console.log(
-          '\n\n<------Matched Action Input in Long Parsing Error------>\n\n',
-          actionInputMatch,
+        logger.debug(
+          '\n\n<------[CustomOutputParser] Matched Action Input in Long Parsing Error------>\n\n' +
+            actionInputMatch,
         );
         return {
           tool: action,
@@ -185,15 +188,14 @@ class CustomOutputParser extends ZeroShotAgentOutputParser {
 
         const inputMatch = this.actionValues.exec(returnValues.log); //new
         if (inputMatch) {
-          console.log('inputMatch');
-          console.dir(inputMatch, { depth: null });
+          logger.debug('[CustomOutputParser] inputMatch', inputMatch);
           returnValues.toolInput = inputMatch[1].replaceAll('"', '').trim();
           returnValues.log = returnValues.log.replace(this.actionValues, '');
         }
 
         return returnValues;
       } else {
-        console.log('No valid tool mentioned.', this.tools, text);
+        logger.debug('[CustomOutputParser] No valid tool mentioned.', this.tools, text);
         return {
           tool: 'self-reflection',
           toolInput: 'Hypothetical actions: \n"' + text + '"\n',
@@ -202,8 +204,8 @@ class CustomOutputParser extends ZeroShotAgentOutputParser {
       }
 
       // if (action && input) {
-      //   console.log('Action:', action);
-      //   console.log('Input:', input);
+      //   logger.debug('Action:', action);
+      //   logger.debug('Input:', input);
       // }
     }
 

api/app/clients/agents/Functions/FunctionsAgent.js

@@ -1,120 +0,0 @@
-const { Agent } = require('langchain/agents');
-const { LLMChain } = require('langchain/chains');
-const { FunctionChatMessage, AIChatMessage } = require('langchain/schema');
-const {
-  ChatPromptTemplate,
-  MessagesPlaceholder,
-  SystemMessagePromptTemplate,
-  HumanMessagePromptTemplate,
-} = require('langchain/prompts');
-const PREFIX = 'You are a helpful AI assistant.';
-
-function parseOutput(message) {
-  if (message.additional_kwargs.function_call) {
-    const function_call = message.additional_kwargs.function_call;
-    return {
-      tool: function_call.name,
-      toolInput: function_call.arguments ? JSON.parse(function_call.arguments) : {},
-      log: message.text,
-    };
-  } else {
-    return { returnValues: { output: message.text }, log: message.text };
-  }
-}
-
-class FunctionsAgent extends Agent {
-  constructor(input) {
-    super({ ...input, outputParser: undefined });
-    this.tools = input.tools;
-  }
-
-  lc_namespace = ['langchain', 'agents', 'openai'];
-
-  _agentType() {
-    return 'openai-functions';
-  }
-
-  observationPrefix() {
-    return 'Observation: ';
-  }
-
-  llmPrefix() {
-    return 'Thought:';
-  }
-
-  _stop() {
-    return ['Observation:'];
-  }
-
-  static createPrompt(_tools, fields) {
-    const { prefix = PREFIX, currentDateString } = fields || {};
-
-    return ChatPromptTemplate.fromMessages([
-      SystemMessagePromptTemplate.fromTemplate(`Date: ${currentDateString}\n${prefix}`),
-      new MessagesPlaceholder('chat_history'),
-      HumanMessagePromptTemplate.fromTemplate('Query: {input}'),
-      new MessagesPlaceholder('agent_scratchpad'),
-    ]);
-  }
-
-  static fromLLMAndTools(llm, tools, args) {
-    FunctionsAgent.validateTools(tools);
-    const prompt = FunctionsAgent.createPrompt(tools, args);
-    const chain = new LLMChain({
-      prompt,
-      llm,
-      callbacks: args?.callbacks,
-    });
-    return new FunctionsAgent({
-      llmChain: chain,
-      allowedTools: tools.map((t) => t.name),
-      tools,
-    });
-  }
-
-  async constructScratchPad(steps) {
-    return steps.flatMap(({ action, observation }) => [
-      new AIChatMessage('', {
-        function_call: {
-          name: action.tool,
-          arguments: JSON.stringify(action.toolInput),
-        },
-      }),
-      new FunctionChatMessage(observation, action.tool),
-    ]);
-  }
-
-  async plan(steps, inputs, callbackManager) {
-    // Add scratchpad and stop to inputs
-    const thoughts = await this.constructScratchPad(steps);
-    const newInputs = Object.assign({}, inputs, { agent_scratchpad: thoughts });
-    if (this._stop().length !== 0) {
-      newInputs.stop = this._stop();
-    }
-
-    // Split inputs between prompt and llm
-    const llm = this.llmChain.llm;
-    const valuesForPrompt = Object.assign({}, newInputs);
-    const valuesForLLM = {
-      tools: this.tools,
-    };
-    for (let i = 0; i < this.llmChain.llm.callKeys.length; i++) {
-      const key = this.llmChain.llm.callKeys[i];
-      if (key in inputs) {
-        valuesForLLM[key] = inputs[key];
-        delete valuesForPrompt[key];
-      }
-    }
-
-    const promptValue = await this.llmChain.prompt.formatPromptValue(valuesForPrompt);
-    const message = await llm.predictMessages(
-      promptValue.toChatMessages(),
-      valuesForLLM,
-      callbackManager,
-    );
-    console.log('message', message);
-    return parseOutput(message);
-  }
-}
-
-module.exports = FunctionsAgent;

api/app/clients/agents/Functions/initializeFunctionsAgent.js

@@ -10,6 +10,8 @@ const initializeFunctionsAgent = async ({
   tools,
   model,
   pastMessages,
+  customName,
+  customInstructions,
   currentDateString,
   ...rest
 }) => {
@@ -24,7 +26,13 @@ const initializeFunctionsAgent = async ({
     returnMessages: true,
   });
 
-  const prefix = addToolDescriptions(`Current Date: ${currentDateString}\n${PREFIX}`, tools);
+  let prefix = addToolDescriptions(`Current Date: ${currentDateString}\n${PREFIX}`, tools);
+  if (customName) {
+    prefix = `You are "${customName}".\n${prefix}`;
+  }
+  if (customInstructions) {
+    prefix = `${prefix}\n${customInstructions}`;
+  }
 
   return await initializeAgentExecutorWithOptions(tools, model, {
     agentType: 'openai-functions',

api/app/clients/callbacks/createStartHandler.js

@@ -1,7 +1,9 @@
 const { promptTokensEstimate } = require('openai-chat-tokens');
-const checkBalance = require('../../../models/checkBalance');
-const { isEnabled } = require('../../../server/utils');
-const { formatFromLangChain } = require('../prompts');
+const { EModelEndpoint, supportsBalanceCheck } = require('librechat-data-provider');
+const { formatFromLangChain } = require('~/app/clients/prompts');
+const checkBalance = require('~/models/checkBalance');
+const { isEnabled } = require('~/server/utils');
+const { logger } = require('~/config');
 
 const createStartHandler = ({
   context,
@@ -15,9 +17,15 @@ const createStartHandler = ({
     const { model, functions, function_call } = invocation_params;
     const messages = _messages[0].map(formatFromLangChain);
 
-    if (manager.debug) {
-      console.log(`handleChatModelStart: ${context}`);
-      console.dir({ model, functions, function_call }, { depth: null });
+    logger.debug(`[createStartHandler] handleChatModelStart: ${context}`, {
+      model,
+      function_call,
+    });
+
+    if (context !== 'title') {
+      logger.debug(`[createStartHandler] handleChatModelStart: ${context}`, {
+        functions,
+      });
     }
 
     const payload = { messages };
@@ -34,13 +42,15 @@ const createStartHandler = ({
     }
 
     prelimPromptTokens += promptTokensEstimate(payload);
-    if (manager.debug) {
-      console.log('Prelim Prompt Tokens & Token Buffer', prelimPromptTokens, tokenBuffer);
-    }
+    logger.debug('[createStartHandler]', {
+      prelimPromptTokens,
+      tokenBuffer,
+    });
     prelimPromptTokens += tokenBuffer;
 
     try {
-      if (isEnabled(process.env.CHECK_BALANCE)) {
+      // TODO: if plugins extends to non-OpenAI models, this will need to be updated
+      if (isEnabled(process.env.CHECK_BALANCE) && supportsBalanceCheck[EModelEndpoint.openAI]) {
         const generations =
           initialMessageCount && messages.length > initialMessageCount
             ? messages.slice(initialMessageCount)
@@ -55,11 +65,12 @@ const createStartHandler = ({
             debug: manager.debug,
             generations,
             model,
+            endpoint: EModelEndpoint.openAI,
           },
         });
       }
     } catch (err) {
-      console.error(`[${context}] checkBalance error`, err);
+      logger.error(`[createStartHandler][${context}] checkBalance error`, err);
       manager.abortController.abort();
       if (context === 'summary' || context === 'plugins') {
         manager.addRun(runId, { conversationId, error: err.message });

api/app/clients/chains/runTitleChain.js

@@ -1,6 +1,7 @@
 const { z } = require('zod');
 const { langPrompt, createTitlePrompt, escapeBraces, getSnippet } = require('../prompts');
 const { createStructuredOutputChainFromZod } = require('langchain/chains/openai_functions');
+const { logger } = require('~/config');
 
 const langSchema = z.object({
   language: z.string().describe('The language of the input text (full noun, no abbreviations).'),
@@ -30,8 +31,7 @@ const runTitleChain = async ({ llm, text, convo, signal, callbacks }) => {
   try {
     snippet = getSnippet(text);
   } catch (e) {
-    console.log('Error getting snippet of text for titleChain');
-    console.log(e);
+    logger.error('[runTitleChain] Error getting snippet of text for titleChain', e);
   }
   const languageChain = createLanguageChain({ llm, callbacks });
   const titleChain = createTitleChain({ llm, callbacks, convo: escapeBraces(convo) });

api/app/clients/document/tokenSplit.js

@@ -1,4 +1,4 @@
-const { TokenTextSplitter } = require('langchain/text_splitter');
+const { TokenTextSplitter } = require('@langchain/textsplitters');
 
 /**
  * Splits a given text by token chunks, based on the provided parameters for the TokenTextSplitter.

api/app/clients/document/tokenSplit.spec.js

@@ -12,7 +12,7 @@ describe('tokenSplit', () => {
       returnSize: 5,
     });
 
-    expect(result).toEqual(['. Null', ' Nullam', 'am id', ' id.', '.']);
+    expect(result).toEqual(['it.', '. Null', ' Nullam', 'am id', ' id.']);
   });
 
   it('returns correct text chunks with default parameters', async () => {

api/app/clients/llm/RunManager.js

@@ -1,5 +1,6 @@
-const { createStartHandler } = require('../callbacks');
-const spendTokens = require('../../../models/spendTokens');
+const { createStartHandler } = require('~/app/clients/callbacks');
+const { spendTokens } = require('~/models/spendTokens');
+const { logger } = require('~/config');
 
 class RunManager {
   constructor(fields) {
@@ -35,7 +36,7 @@ class RunManager {
     if (this.runs.has(runId)) {
       this.runs.delete(runId);
     } else {
-      console.error(`Run with ID ${runId} does not exist.`);
+      logger.error(`[api/app/clients/llm/RunManager] Run with ID ${runId} does not exist.`);
     }
   }
 
@@ -57,10 +58,19 @@ class RunManager {
       {
         handleChatModelStart: createStartHandler({ ...metadata, manager: this }),
         handleLLMEnd: async (output, runId, _parentRunId) => {
-          if (this.debug) {
-            console.log(`handleLLMEnd: ${JSON.stringify(metadata)}`);
-            console.dir({ output, runId, _parentRunId }, { depth: null });
+          const { llmOutput, ..._output } = output;
+          logger.debug(`[RunManager] handleLLMEnd: ${JSON.stringify(metadata)}`, {
+            runId,
+            _parentRunId,
+            llmOutput,
+          });
+
+          if (metadata.context !== 'title') {
+            logger.debug('[RunManager] handleLLMEnd:', {
+              output: _output,
+            });
           }
+
           const { tokenUsage } = output.llmOutput;
           const run = this.getRunById(runId);
           this.removeRun(runId);
@@ -74,8 +84,7 @@ class RunManager {
           await spendTokens(txData, tokenUsage);
         },
         handleLLMError: async (err) => {
-          this.debug && console.log(`handleLLMError: ${JSON.stringify(metadata)}`);
-          this.debug && console.error(err);
+          logger.error(`[RunManager] handleLLMError: ${JSON.stringify(metadata)}`, err);
           if (metadata.context === 'title') {
             return;
           } else if (metadata.context === 'plugins') {

api/app/clients/llm/createCoherePayload.js

@@ -0,0 +1,85 @@
+const { CohereConstants } = require('librechat-data-provider');
+const { titleInstruction } = require('../prompts/titlePrompts');
+
+// Mapping OpenAI roles to Cohere roles
+const roleMap = {
+  user: CohereConstants.ROLE_USER,
+  assistant: CohereConstants.ROLE_CHATBOT,
+  system: CohereConstants.ROLE_SYSTEM, // Recognize and map the system role explicitly
+};
+
+/**
+ * Adjusts an OpenAI ChatCompletionPayload to conform with Cohere's expected chat payload format.
+ * Now includes handling for "system" roles explicitly mentioned.
+ *
+ * @param {Object} options - Object containing the model options.
+ * @param {ChatCompletionPayload} options.modelOptions - The OpenAI model payload options.
+ * @returns {CohereChatStreamRequest} Cohere-compatible chat API payload.
+ */
+function createCoherePayload({ modelOptions }) {
+  /** @type {string | undefined} */
+  let preamble;
+  let latestUserMessageContent = '';
+  const {
+    stream,
+    stop,
+    top_p,
+    temperature,
+    frequency_penalty,
+    presence_penalty,
+    max_tokens,
+    messages,
+    model,
+    ...rest
+  } = modelOptions;
+
+  // Filter out the latest user message and transform remaining messages to Cohere's chat_history format
+  let chatHistory = messages.reduce((acc, message, index, arr) => {
+    const isLastUserMessage = index === arr.length - 1 && message.role === 'user';
+
+    const messageContent =
+      typeof message.content === 'string'
+        ? message.content
+        : message.content.map((part) => (part.type === 'text' ? part.text : '')).join(' ');
+
+    if (isLastUserMessage) {
+      latestUserMessageContent = messageContent;
+    } else {
+      acc.push({
+        role: roleMap[message.role] || CohereConstants.ROLE_USER,
+        message: messageContent,
+      });
+    }
+
+    return acc;
+  }, []);
+
+  if (
+    chatHistory.length === 1 &&
+    chatHistory[0].role === CohereConstants.ROLE_SYSTEM &&
+    !latestUserMessageContent.length
+  ) {
+    const message = chatHistory[0].message;
+    latestUserMessageContent = message.includes(titleInstruction)
+      ? CohereConstants.TITLE_MESSAGE
+      : '.';
+    preamble = message;
+  }
+
+  return {
+    message: latestUserMessageContent,
+    model: model,
+    chatHistory,
+    stream: stream ?? false,
+    temperature: temperature,
+    frequencyPenalty: frequency_penalty,
+    presencePenalty: presence_penalty,
+    maxTokens: max_tokens,
+    stopSequences: stop,
+    preamble,
+    p: top_p,
+    ...rest,
+  };
+}
+
+module.exports = createCoherePayload;

api/app/clients/llm/createLLM.js

@@ -1,38 +1,6 @@
-const { ChatOpenAI } = require('langchain/chat_models/openai');
-const { sanitizeModelName } = require('../../../utils');
-const { isEnabled } = require('../../../server/utils');
-
-/**
- * @typedef {Object} ModelOptions
- * @property {string} modelName - The name of the model.
- * @property {number} [temperature] - The temperature setting for the model.
- * @property {number} [presence_penalty] - The presence penalty setting.
- * @property {number} [frequency_penalty] - The frequency penalty setting.
- * @property {number} [max_tokens] - The maximum number of tokens to generate.
- */
-
-/**
- * @typedef {Object} ConfigOptions
- * @property {string} [basePath] - The base path for the API requests.
- * @property {Object} [baseOptions] - Base options for the API requests, including headers.
- * @property {Object} [httpAgent] - The HTTP agent for the request.
- * @property {Object} [httpsAgent] - The HTTPS agent for the request.
- */
-
-/**
- * @typedef {Object} Callbacks
- * @property {Function} [handleChatModelStart] - A callback function for handleChatModelStart
- * @property {Function} [handleLLMEnd] - A callback function for handleLLMEnd
- * @property {Function} [handleLLMError] - A callback function for handleLLMError
- */
-
-/**
- * @typedef {Object} AzureOptions
- * @property {string} [azureOpenAIApiKey] - The Azure OpenAI API key.
- * @property {string} [azureOpenAIApiInstanceName] - The Azure OpenAI API instance name.
- * @property {string} [azureOpenAIApiDeploymentName] - The Azure OpenAI API deployment name.
- * @property {string} [azureOpenAIApiVersion] - The Azure OpenAI API version.
- */
+const { ChatOpenAI } = require('@langchain/openai');
+const { sanitizeModelName, constructAzureURL } = require('~/utils');
+const { isEnabled } = require('~/server/utils');
 
 /**
  * Creates a new instance of a language model (LLM) for chat interactions.
@@ -40,7 +8,7 @@ const { isEnabled } = require('../../../server/utils');
  * @param {Object} options - The options for creating the LLM.
  * @param {ModelOptions} options.modelOptions - The options specific to the model, including modelName, temperature, presence_penalty, frequency_penalty, and other model-related settings.
  * @param {ConfigOptions} options.configOptions - Configuration options for the API requests, including proxy settings and custom headers.
- * @param {Callbacks} options.callbacks - Callback functions for managing the lifecycle of the LLM, including token buffers, context, and initial message count.
+ * @param {Callbacks} [options.callbacks] - Callback functions for managing the lifecycle of the LLM, including token buffers, context, and initial message count.
  * @param {boolean} [options.streaming=false] - Determines if the LLM should operate in streaming mode.
  * @param {string} options.openAIApiKey - The API key for OpenAI, used for authentication.
  * @param {AzureOptions} [options.azure={}] - Optional Azure-specific configurations. If provided, Azure configurations take precedence over OpenAI configurations.
@@ -49,7 +17,7 @@ const { isEnabled } = require('../../../server/utils');
  *
  * @example
  * const llm = createLLM({
- *   modelOptions: { modelName: 'gpt-3.5-turbo', temperature: 0.2 },
+ *   modelOptions: { modelName: 'gpt-4o-mini', temperature: 0.2 },
  *   configOptions: { basePath: 'https://example.api/path' },
  *   callbacks: { onMessage: handleMessage },
  *   openAIApiKey: 'your-api-key'
@@ -68,6 +36,7 @@ function createLLM({
     apiKey: openAIApiKey,
   };
 
+  /**  @type {AzureOptions} */
   let azureOptions = {};
   if (azure) {
     const useModelName = isEnabled(process.env.AZURE_USE_MODEL_AS_DEPLOYMENT_NAME);
@@ -85,17 +54,24 @@ function createLLM({
     modelOptions.modelName = process.env.AZURE_OPENAI_DEFAULT_MODEL;
   }
 
-  // console.debug('createLLM: configOptions');
-  // console.debug(configOptions);
+  if (azure && configOptions.basePath) {
+    const azureURL = constructAzureURL({
+      baseURL: configOptions.basePath,
+      azureOptions,
+    });
+    azureOptions.azureOpenAIBasePath = azureURL.split(
+      `/${azureOptions.azureOpenAIApiDeploymentName}`,
+    )[0];
+  }
 
   return new ChatOpenAI(
     {
       streaming,
-      verbose: true,
       credentials,
       configuration,
       ...azureOptions,
       ...modelOptions,
+      ...credentials,
       callbacks,
     },
     configOptions,

api/app/clients/llm/index.js

@@ -1,7 +1,9 @@
 const createLLM = require('./createLLM');
 const RunManager = require('./RunManager');
+const createCoherePayload = require('./createCoherePayload');
 
 module.exports = {
   createLLM,
   RunManager,
+  createCoherePayload,
 };

api/app/clients/memory/summaryBuffer.demo.js

@@ -1,9 +1,9 @@
 require('dotenv').config();
-const { ChatOpenAI } = require('langchain/chat_models/openai');
+const { ChatOpenAI } = require('@langchain/openai');
 const { getBufferString, ConversationSummaryBufferMemory } = require('langchain/memory');
 
 const chatPromptMemory = new ConversationSummaryBufferMemory({
-  llm: new ChatOpenAI({ modelName: 'gpt-3.5-turbo', temperature: 0 }),
+  llm: new ChatOpenAI({ modelName: 'gpt-4o-mini', temperature: 0 }),
   maxTokenLimit: 10,
   returnMessages: true,
 });

api/app/clients/memory/summaryBuffer.js

@@ -1,6 +1,7 @@
 const { ConversationSummaryBufferMemory, ChatMessageHistory } = require('langchain/memory');
 const { formatLangChainMessages, SUMMARY_PROMPT } = require('../prompts');
 const { predictNewSummary } = require('../chains');
+const { logger } = require('~/config');
 
 const createSummaryBufferMemory = ({ llm, prompt, messages, ...rest }) => {
   const chatHistory = new ChatMessageHistory(messages);
@@ -22,9 +23,8 @@ const summaryBuffer = async ({
   prompt = SUMMARY_PROMPT,
   signal,
 }) => {
-  if (debug && previous_summary) {
-    console.log('<-----------PREVIOUS SUMMARY----------->\n\n');
-    console.log(previous_summary);
+  if (previous_summary) {
+    logger.debug('[summaryBuffer]', { previous_summary });
   }
 
   const formattedMessages = formatLangChainMessages(context, formatOptions);
@@ -46,8 +46,7 @@ const summaryBuffer = async ({
   const messages = await chatPromptMemory.chatHistory.getMessages();
 
   if (debug) {
-    console.log('<-----------SUMMARY BUFFER MESSAGES----------->\n\n');
-    console.log(JSON.stringify(messages));
+    logger.debug('[summaryBuffer]', { summary_buffer_messages: messages.length });
   }
 
   const predictSummary = await predictNewSummary({
@@ -58,8 +57,7 @@ const summaryBuffer = async ({
   });
 
   if (debug) {
-    console.log('<-----------SUMMARY----------->\n\n');
-    console.log(JSON.stringify(predictSummary));
+    logger.debug('[summaryBuffer]', { summary: predictSummary });
   }
 
   return { role: 'system', content: predictSummary };

api/app/clients/output_parsers/addImages.js

@@ -1,3 +1,5 @@
+const { logger } = require('~/config');
+
 /**
  * The `addImages` function corrects any erroneous image URLs in the `responseMessage.text`
  * and appends image observations from `intermediateSteps` if they are not already present.
@@ -20,7 +22,7 @@
  *
  * addImages(intermediateSteps, responseMessage);
  *
- * console.log(responseMessage.text);
+ * logger.debug(responseMessage.text);
  * // Outputs: 'Some text with ![desc](/images/test.png)\n![desc](/images/test.png)'
  *
  * @returns {void}
@@ -58,12 +60,10 @@ function addImages(intermediateSteps, responseMessage) {
     if (!observation || !observation.includes('![')) {
       return;
     }
-    const observedImagePath = observation.match(/\(\/images\/.*\.\w*\)/g);
+    const observedImagePath = observation.match(/!\[[^(]*\]\([^)]*\)/g);
     if (observedImagePath && !responseMessage.text.includes(observedImagePath[0])) {
-      responseMessage.text += '\n' + observation;
-      if (process.env.DEBUG_PLUGINS) {
-        console.debug('[addImages] added image from intermediateSteps');
-      }
+      responseMessage.text += '\n' + observedImagePath[0];
+      logger.debug('[addImages] added image from intermediateSteps:', observedImagePath[0]);
     }
   });
 }

api/app/clients/output_parsers/addImages.spec.js

@@ -81,4 +81,62 @@ describe('addImages', () => {
     addImages(intermediateSteps, responseMessage);
     expect(responseMessage.text).toBe(`${originalText}\n${imageMarkdown}`);
   });
+
+  it('should extract only image markdowns when there is text between them', () => {
+    const markdownWithTextBetweenImages = `
+      ![image1](/images/image1.png)
+      Some text between images that should not be included.
+      ![image2](/images/image2.png)
+      More text that should be ignored.
+      ![image3](/images/image3.png)
+    `;
+    intermediateSteps.push({ observation: markdownWithTextBetweenImages });
+    addImages(intermediateSteps, responseMessage);
+    expect(responseMessage.text).toBe('\n![image1](/images/image1.png)');
+  });
+
+  it('should only return the first image when multiple images are present', () => {
+    const markdownWithMultipleImages = `
+      ![image1](/images/image1.png)
+      ![image2](/images/image2.png)
+      ![image3](/images/image3.png)
+    `;
+    intermediateSteps.push({ observation: markdownWithMultipleImages });
+    addImages(intermediateSteps, responseMessage);
+    expect(responseMessage.text).toBe('\n![image1](/images/image1.png)');
+  });
+
+  it('should not include any text or metadata surrounding the image markdown', () => {
+    const markdownWithMetadata = `
+      Title: Test Document
+      Author: John Doe
+      ![image1](/images/image1.png)
+      Some content after the image.
+      Vector values: [0.1, 0.2, 0.3]
+    `;
+    intermediateSteps.push({ observation: markdownWithMetadata });
+    addImages(intermediateSteps, responseMessage);
+    expect(responseMessage.text).toBe('\n![image1](/images/image1.png)');
+  });
+
+  it('should handle complex markdown with multiple images and only return the first one', () => {
+    const complexMarkdown = `
+      # Document Title
+      
+      ## Section 1
+      Here's some text with an embedded image:
+      ![image1](/images/image1.png)
+      
+      ## Section 2
+      More text here...
+      ![image2](/images/image2.png)
+      
+      ### Subsection
+      Even more content
+      ![image3](/images/image3.png)
+    `;
+    intermediateSteps.push({ observation: complexMarkdown });
+    addImages(intermediateSteps, responseMessage);
+    expect(responseMessage.text).toBe('\n![image1](/images/image1.png)');
+  });
 });

api/app/clients/prompts/addCacheControl.js

@@ -0,0 +1,45 @@
+/**
+ * Anthropic API: Adds cache control to the appropriate user messages in the payload.
+ * @param {Array<AnthropicMessage | BaseMessage>} messages - The array of message objects.
+ * @returns {Array<AnthropicMessage | BaseMessage>} - The updated array of message objects with cache control added.
+ */
+function addCacheControl(messages) {
+  if (!Array.isArray(messages) || messages.length < 2) {
+    return messages;
+  }
+
+  const updatedMessages = [...messages];
+  let userMessagesModified = 0;
+
+  for (let i = updatedMessages.length - 1; i >= 0 && userMessagesModified < 2; i--) {
+    const message = updatedMessages[i];
+    if (message.getType != null && message.getType() !== 'human') {
+      continue;
+    } else if (message.getType == null && message.role !== 'user') {
+      continue;
+    }
+
+    if (typeof message.content === 'string') {
+      message.content = [
+        {
+          type: 'text',
+          text: message.content,
+          cache_control: { type: 'ephemeral' },
+        },
+      ];
+      userMessagesModified++;
+    } else if (Array.isArray(message.content)) {
+      for (let j = message.content.length - 1; j >= 0; j--) {
+        if (message.content[j].type === 'text') {
+          message.content[j].cache_control = { type: 'ephemeral' };
+          userMessagesModified++;
+          break;
+        }
+      }
+    }
+  }
+
+  return updatedMessages;
+}
+
+module.exports = addCacheControl;

api/app/clients/prompts/addCacheControl.spec.js

@@ -0,0 +1,227 @@
+const addCacheControl = require('./addCacheControl');
+
+describe('addCacheControl', () => {
+  test('should add cache control to the last two user messages with array content', () => {
+    const messages = [
+      { role: 'user', content: [{ type: 'text', text: 'Hello' }] },
+      { role: 'assistant', content: [{ type: 'text', text: 'Hi there' }] },
+      { role: 'user', content: [{ type: 'text', text: 'How are you?' }] },
+      { role: 'assistant', content: [{ type: 'text', text: 'I\'m doing well, thanks!' }] },
+      { role: 'user', content: [{ type: 'text', text: 'Great!' }] },
+    ];
+
+    const result = addCacheControl(messages);
+
+    expect(result[0].content[0]).not.toHaveProperty('cache_control');
+    expect(result[2].content[0].cache_control).toEqual({ type: 'ephemeral' });
+    expect(result[4].content[0].cache_control).toEqual({ type: 'ephemeral' });
+  });
+
+  test('should add cache control to the last two user messages with string content', () => {
+    const messages = [
+      { role: 'user', content: 'Hello' },
+      { role: 'assistant', content: 'Hi there' },
+      { role: 'user', content: 'How are you?' },
+      { role: 'assistant', content: 'I\'m doing well, thanks!' },
+      { role: 'user', content: 'Great!' },
+    ];
+
+    const result = addCacheControl(messages);
+
+    expect(result[0].content).toBe('Hello');
+    expect(result[2].content[0]).toEqual({
+      type: 'text',
+      text: 'How are you?',
+      cache_control: { type: 'ephemeral' },
+    });
+    expect(result[4].content[0]).toEqual({
+      type: 'text',
+      text: 'Great!',
+      cache_control: { type: 'ephemeral' },
+    });
+  });
+
+  test('should handle mixed string and array content', () => {
+    const messages = [
+      { role: 'user', content: 'Hello' },
+      { role: 'assistant', content: 'Hi there' },
+      { role: 'user', content: [{ type: 'text', text: 'How are you?' }] },
+    ];
+
+    const result = addCacheControl(messages);
+
+    expect(result[0].content[0]).toEqual({
+      type: 'text',
+      text: 'Hello',
+      cache_control: { type: 'ephemeral' },
+    });
+    expect(result[2].content[0].cache_control).toEqual({ type: 'ephemeral' });
+  });
+
+  test('should handle less than two user messages', () => {
+    const messages = [
+      { role: 'user', content: 'Hello' },
+      { role: 'assistant', content: 'Hi there' },
+    ];
+
+    const result = addCacheControl(messages);
+
+    expect(result[0].content[0]).toEqual({
+      type: 'text',
+      text: 'Hello',
+      cache_control: { type: 'ephemeral' },
+    });
+    expect(result[1].content).toBe('Hi there');
+  });
+
+  test('should return original array if no user messages', () => {
+    const messages = [
+      { role: 'assistant', content: 'Hi there' },
+      { role: 'assistant', content: 'How can I help?' },
+    ];
+
+    const result = addCacheControl(messages);
+
+    expect(result).toEqual(messages);
+  });
+
+  test('should handle empty array', () => {
+    const messages = [];
+    const result = addCacheControl(messages);
+    expect(result).toEqual([]);
+  });
+
+  test('should handle non-array input', () => {
+    const messages = 'not an array';
+    const result = addCacheControl(messages);
+    expect(result).toBe('not an array');
+  });
+
+  test('should not modify assistant messages', () => {
+    const messages = [
+      { role: 'user', content: 'Hello' },
+      { role: 'assistant', content: 'Hi there' },
+      { role: 'user', content: 'How are you?' },
+    ];
+
+    const result = addCacheControl(messages);
+
+    expect(result[1].content).toBe('Hi there');
+  });
+
+  test('should handle multiple content items in user messages', () => {
+    const messages = [
+      {
+        role: 'user',
+        content: [
+          { type: 'text', text: 'Hello' },
+          { type: 'image', url: 'http://example.com/image.jpg' },
+          { type: 'text', text: 'This is an image' },
+        ],
+      },
+      { role: 'assistant', content: 'Hi there' },
+      { role: 'user', content: 'How are you?' },
+    ];
+
+    const result = addCacheControl(messages);
+
+    expect(result[0].content[0]).not.toHaveProperty('cache_control');
+    expect(result[0].content[1]).not.toHaveProperty('cache_control');
+    expect(result[0].content[2].cache_control).toEqual({ type: 'ephemeral' });
+    expect(result[2].content[0]).toEqual({
+      type: 'text',
+      text: 'How are you?',
+      cache_control: { type: 'ephemeral' },
+    });
+  });
+
+  test('should handle an array with mixed content types', () => {
+    const messages = [
+      { role: 'user', content: 'Hello' },
+      { role: 'assistant', content: 'Hi there' },
+      { role: 'user', content: [{ type: 'text', text: 'How are you?' }] },
+      { role: 'assistant', content: 'I\'m doing well, thanks!' },
+      { role: 'user', content: 'Great!' },
+    ];
+
+    const result = addCacheControl(messages);
+
+    expect(result[0].content).toEqual('Hello');
+    expect(result[2].content[0]).toEqual({
+      type: 'text',
+      text: 'How are you?',
+      cache_control: { type: 'ephemeral' },
+    });
+    expect(result[4].content).toEqual([
+      {
+        type: 'text',
+        text: 'Great!',
+        cache_control: { type: 'ephemeral' },
+      },
+    ]);
+    expect(result[1].content).toBe('Hi there');
+    expect(result[3].content).toBe('I\'m doing well, thanks!');
+  });
+
+  test('should handle edge case with multiple content types', () => {
+    const messages = [
+      {
+        role: 'user',
+        content: [
+          {
+            type: 'image',
+            source: { type: 'base64', media_type: 'image/png', data: 'some_base64_string' },
+          },
+          {
+            type: 'image',
+            source: { type: 'base64', media_type: 'image/png', data: 'another_base64_string' },
+          },
+          { type: 'text', text: 'what do all these images have in common' },
+        ],
+      },
+      { role: 'assistant', content: 'I see multiple images.' },
+      { role: 'user', content: 'Correct!' },
+    ];
+
+    const result = addCacheControl(messages);
+
+    expect(result[0].content[0]).not.toHaveProperty('cache_control');
+    expect(result[0].content[1]).not.toHaveProperty('cache_control');
+    expect(result[0].content[2].cache_control).toEqual({ type: 'ephemeral' });
+    expect(result[2].content[0]).toEqual({
+      type: 'text',
+      text: 'Correct!',
+      cache_control: { type: 'ephemeral' },
+    });
+  });
+
+  test('should handle user message with no text block', () => {
+    const messages = [
+      {
+        role: 'user',
+        content: [
+          {
+            type: 'image',
+            source: { type: 'base64', media_type: 'image/png', data: 'some_base64_string' },
+          },
+          {
+            type: 'image',
+            source: { type: 'base64', media_type: 'image/png', data: 'another_base64_string' },
+          },
+        ],
+      },
+      { role: 'assistant', content: 'I see two images.' },
+      { role: 'user', content: 'Correct!' },
+    ];
+
+    const result = addCacheControl(messages);
+
+    expect(result[0].content[0]).not.toHaveProperty('cache_control');
+    expect(result[0].content[1]).not.toHaveProperty('cache_control');
+    expect(result[2].content[0]).toEqual({
+      type: 'text',
+      text: 'Correct!',
+      cache_control: { type: 'ephemeral' },
+    });
+  });
+});

api/app/clients/prompts/artifacts.js

@@ -0,0 +1,527 @@
+const dedent = require('dedent');
+const { EModelEndpoint, ArtifactModes } = require('librechat-data-provider');
+const { generateShadcnPrompt } = require('~/app/clients/prompts/shadcn-docs/generate');
+const { components } = require('~/app/clients/prompts/shadcn-docs/components');
+
+// eslint-disable-next-line no-unused-vars
+const artifactsPromptV1 = dedent`The assistant can create and reference artifacts during conversations.
+  
+Artifacts are for substantial, self-contained content that users might modify or reuse, displayed in a separate UI window for clarity.
+
+# Good artifacts are...
+- Substantial content (>15 lines)
+- Content that the user is likely to modify, iterate on, or take ownership of
+- Self-contained, complex content that can be understood on its own, without context from the conversation
+- Content intended for eventual use outside the conversation (e.g., reports, emails, presentations)
+- Content likely to be referenced or reused multiple times
+
+# Don't use artifacts for...
+- Simple, informational, or short content, such as brief code snippets, mathematical equations, or small examples
+- Primarily explanatory, instructional, or illustrative content, such as examples provided to clarify a concept
+- Suggestions, commentary, or feedback on existing artifacts
+- Conversational or explanatory content that doesn't represent a standalone piece of work
+- Content that is dependent on the current conversational context to be useful
+- Content that is unlikely to be modified or iterated upon by the user
+- Request from users that appears to be a one-off question
+
+# Usage notes
+- One artifact per message unless specifically requested
+- Prefer in-line content (don't use artifacts) when possible. Unnecessary use of artifacts can be jarring for users.
+- If a user asks the assistant to "draw an SVG" or "make a website," the assistant does not need to explain that it doesn't have these capabilities. Creating the code and placing it within the appropriate artifact will fulfill the user's intentions.
+- If asked to generate an image, the assistant can offer an SVG instead. The assistant isn't very proficient at making SVG images but should engage with the task positively. Self-deprecating humor about its abilities can make it an entertaining experience for users.
+- The assistant errs on the side of simplicity and avoids overusing artifacts for content that can be effectively presented within the conversation.
+- Always provide complete, specific, and fully functional content without any placeholders, ellipses, or 'remains the same' comments.
+
+<artifact_instructions>
+  When collaborating with the user on creating content that falls into compatible categories, the assistant should follow these steps:
+
+  1. Create the artifact using the following format:
+
+     :::artifact{identifier="unique-identifier" type="mime-type" title="Artifact Title"}
+     \`\`\`
+     Your artifact content here
+     \`\`\`
+     :::
+
+  2. Assign an identifier to the \`identifier\` attribute. For updates, reuse the prior identifier. For new artifacts, the identifier should be descriptive and relevant to the content, using kebab-case (e.g., "example-code-snippet"). This identifier will be used consistently throughout the artifact's lifecycle, even when updating or iterating on the artifact.
+  3. Include a \`title\` attribute to provide a brief title or description of the content.
+  4. Add a \`type\` attribute to specify the type of content the artifact represents. Assign one of the following values to the \`type\` attribute:
+    - HTML: "text/html"
+      - The user interface can render single file HTML pages placed within the artifact tags. HTML, JS, and CSS should be in a single file when using the \`text/html\` type.
+      - Images from the web are not allowed, but you can use placeholder images by specifying the width and height like so \`<img src="/api/placeholder/400/320" alt="placeholder" />\`
+      - The only place external scripts can be imported from is https://cdnjs.cloudflare.com
+    - Mermaid Diagrams: "application/vnd.mermaid"
+      - The user interface will render Mermaid diagrams placed within the artifact tags.
+    - React Components: "application/vnd.react"
+      - Use this for displaying either: React elements, e.g. \`<strong>Hello World!</strong>\`, React pure functional components, e.g. \`() => <strong>Hello World!</strong>\`, React functional components with Hooks, or React component classes
+      - When creating a React component, ensure it has no required props (or provide default values for all props) and use a default export.
+      - Use Tailwind classes for styling. DO NOT USE ARBITRARY VALUES (e.g. \`h-[600px]\`).
+      - Base React is available to be imported. To use hooks, first import it at the top of the artifact, e.g. \`import { useState } from "react"\`
+      - The lucide-react@0.263.1 library is available to be imported. e.g. \`import { Camera } from "lucide-react"\` & \`<Camera color="red" size={48} />\`
+      - The recharts charting library is available to be imported, e.g. \`import { LineChart, XAxis, ... } from "recharts"\` & \`<LineChart ...><XAxis dataKey="name"> ...\`
+      - The assistant can use prebuilt components from the \`shadcn/ui\` library after it is imported: \`import { Alert, AlertDescription, AlertTitle, AlertDialog, AlertDialogAction } from '/components/ui/alert';\`. If using components from the shadcn/ui library, the assistant mentions this to the user and offers to help them install the components if necessary.
+      - Components MUST be imported from \`/components/ui/name\` and NOT from \`/components/name\` or \`@/components/ui/name\`.
+      - NO OTHER LIBRARIES (e.g. zod, hookform) ARE INSTALLED OR ABLE TO BE IMPORTED.
+      - Images from the web are not allowed, but you can use placeholder images by specifying the width and height like so \`<img src="/api/placeholder/400/320" alt="placeholder" />\`
+      - If you are unable to follow the above requirements for any reason, don't use artifacts and use regular code blocks instead, which will not attempt to render the component.
+  5. Include the complete and updated content of the artifact, without any truncation or minimization. Don't use "// rest of the code remains the same...".
+  6. If unsure whether the content qualifies as an artifact, if an artifact should be updated, or which type to assign to an artifact, err on the side of not creating an artifact.
+  7. Always use triple backticks (\`\`\`) to enclose the content within the artifact, regardless of the content type.
+</artifact_instructions>
+
+Here are some examples of correct usage of artifacts:
+
+<examples>
+  <example_docstring>
+    This example demonstrates how to create a Mermaid artifact for a simple flow chart.
+  </example_docstring>
+
+  <example>
+    <user_query>Can you create a simple flow chart showing the process of making tea using Mermaid?</user_query>
+
+    <assistant_response>
+      Sure! Here's a simple flow chart depicting the process of making tea using Mermaid syntax:
+
+      :::artifact{identifier="tea-making-flowchart" type="application/vnd.mermaid" title="Flow chart: Making Tea"}
+      \`\`\`mermaid
+      graph TD
+          A[Start] --> B{Water boiled?}
+          B -->|Yes| C[Add tea leaves to cup]
+          B -->|No| D[Boil water]
+          D --> B
+          C --> E[Pour boiling water into cup]
+          E --> F[Steep tea for desired time]
+          F --> G[Remove tea leaves]
+          G --> H[Add milk or sugar, if desired]
+          H --> I[Enjoy your tea!]
+          I --> J[End]
+      \`\`\`
+      :::
+
+      This flow chart uses Mermaid syntax to visualize the steps involved in making a cup of tea. Here's a brief explanation of the process:
+
+      1. Start
+      2. Check if water is boiled
+      3. If not boiled, boil the water
+      4. Once water is boiled, add tea leaves to the cup
+      5. Pour boiling water into the cup
+      6. Steep the tea for the desired time
+      7. Remove the tea leaves
+      8. Optionally add milk or sugar
+      9. Enjoy your tea!
+      10. End
+
+      This chart provides a clear visual representation of the tea-making process. You can easily modify or expand this chart if you want to add more details or steps to the process. Let me know if you'd like any changes or have any questions!
+    </assistant_response>
+  </example>
+</examples>`;
+const artifactsPrompt = dedent`The assistant can create and reference artifacts during conversations.
+  
+Artifacts are for substantial, self-contained content that users might modify or reuse, displayed in a separate UI window for clarity.
+
+# Good artifacts are...
+- Substantial content (>15 lines)
+- Content that the user is likely to modify, iterate on, or take ownership of
+- Self-contained, complex content that can be understood on its own, without context from the conversation
+- Content intended for eventual use outside the conversation (e.g., reports, emails, presentations)
+- Content likely to be referenced or reused multiple times
+
+# Don't use artifacts for...
+- Simple, informational, or short content, such as brief code snippets, mathematical equations, or small examples
+- Primarily explanatory, instructional, or illustrative content, such as examples provided to clarify a concept
+- Suggestions, commentary, or feedback on existing artifacts
+- Conversational or explanatory content that doesn't represent a standalone piece of work
+- Content that is dependent on the current conversational context to be useful
+- Content that is unlikely to be modified or iterated upon by the user
+- Request from users that appears to be a one-off question
+
+# Usage notes
+- One artifact per message unless specifically requested
+- Prefer in-line content (don't use artifacts) when possible. Unnecessary use of artifacts can be jarring for users.
+- If a user asks the assistant to "draw an SVG" or "make a website," the assistant does not need to explain that it doesn't have these capabilities. Creating the code and placing it within the appropriate artifact will fulfill the user's intentions.
+- If asked to generate an image, the assistant can offer an SVG instead. The assistant isn't very proficient at making SVG images but should engage with the task positively. Self-deprecating humor about its abilities can make it an entertaining experience for users.
+- The assistant errs on the side of simplicity and avoids overusing artifacts for content that can be effectively presented within the conversation.
+- Always provide complete, specific, and fully functional content for artifacts without any snippets, placeholders, ellipses, or 'remains the same' comments.
+- If an artifact is not necessary or requested, the assistant should not mention artifacts at all, and respond to the user accordingly.
+
+<artifact_instructions>
+  When collaborating with the user on creating content that falls into compatible categories, the assistant should follow these steps:
+
+  1. Create the artifact using the following format:
+
+     :::artifact{identifier="unique-identifier" type="mime-type" title="Artifact Title"}
+     \`\`\`
+     Your artifact content here
+     \`\`\`
+     :::
+
+  2. Assign an identifier to the \`identifier\` attribute. For updates, reuse the prior identifier. For new artifacts, the identifier should be descriptive and relevant to the content, using kebab-case (e.g., "example-code-snippet"). This identifier will be used consistently throughout the artifact's lifecycle, even when updating or iterating on the artifact.
+  3. Include a \`title\` attribute to provide a brief title or description of the content.
+  4. Add a \`type\` attribute to specify the type of content the artifact represents. Assign one of the following values to the \`type\` attribute:
+    - HTML: "text/html"
+      - The user interface can render single file HTML pages placed within the artifact tags. HTML, JS, and CSS should be in a single file when using the \`text/html\` type.
+      - Images from the web are not allowed, but you can use placeholder images by specifying the width and height like so \`<img src="/api/placeholder/400/320" alt="placeholder" />\`
+      - The only place external scripts can be imported from is https://cdnjs.cloudflare.com
+    - SVG: "image/svg+xml"
+      - The user interface will render the Scalable Vector Graphics (SVG) image within the artifact tags.
+      - The assistant should specify the viewbox of the SVG rather than defining a width/height
+    - Mermaid Diagrams: "application/vnd.mermaid"
+      - The user interface will render Mermaid diagrams placed within the artifact tags.
+    - React Components: "application/vnd.react"
+      - Use this for displaying either: React elements, e.g. \`<strong>Hello World!</strong>\`, React pure functional components, e.g. \`() => <strong>Hello World!</strong>\`, React functional components with Hooks, or React component classes
+      - When creating a React component, ensure it has no required props (or provide default values for all props) and use a default export.
+      - Use Tailwind classes for styling. DO NOT USE ARBITRARY VALUES (e.g. \`h-[600px]\`).
+      - Base React is available to be imported. To use hooks, first import it at the top of the artifact, e.g. \`import { useState } from "react"\`
+      - The lucide-react@0.394.0 library is available to be imported. e.g. \`import { Camera } from "lucide-react"\` & \`<Camera color="red" size={48} />\`
+      - The recharts charting library is available to be imported, e.g. \`import { LineChart, XAxis, ... } from "recharts"\` & \`<LineChart ...><XAxis dataKey="name"> ...\`
+      - The three.js library is available to be imported, e.g. \`import * as THREE from "three";\`
+      - The date-fns library is available to be imported, e.g. \`import { compareAsc, format } from "date-fns";\`
+      - The react-day-picker library is available to be imported, e.g. \`import { DayPicker } from "react-day-picker";\`
+      - The assistant can use prebuilt components from the \`shadcn/ui\` library after it is imported: \`import { Alert, AlertDescription, AlertTitle, AlertDialog, AlertDialogAction } from '/components/ui/alert';\`. If using components from the shadcn/ui library, the assistant mentions this to the user and offers to help them install the components if necessary.
+      - Components MUST be imported from \`/components/ui/name\` and NOT from \`/components/name\` or \`@/components/ui/name\`.
+      - NO OTHER LIBRARIES (e.g. zod, hookform) ARE INSTALLED OR ABLE TO BE IMPORTED.
+      - Images from the web are not allowed, but you can use placeholder images by specifying the width and height like so \`<img src="/api/placeholder/400/320" alt="placeholder" />\`
+      - When iterating on code, ensure that the code is complete and functional without any snippets, placeholders, or ellipses.
+      - If you are unable to follow the above requirements for any reason, don't use artifacts and use regular code blocks instead, which will not attempt to render the component.
+  5. Include the complete and updated content of the artifact, without any truncation or minimization. Don't use "// rest of the code remains the same...".
+  6. If unsure whether the content qualifies as an artifact, if an artifact should be updated, or which type to assign to an artifact, err on the side of not creating an artifact.
+  7. Always use triple backticks (\`\`\`) to enclose the content within the artifact, regardless of the content type.
+</artifact_instructions>
+
+Here are some examples of correct usage of artifacts:
+
+<examples>
+  <example_docstring>
+    This example demonstrates how to create a Mermaid artifact for a simple flow chart.
+  </example_docstring>
+
+  <example>
+    <user_query>Can you create a simple flow chart showing the process of making tea using Mermaid?</user_query>
+
+    <assistant_response>
+      Sure! Here's a simple flow chart depicting the process of making tea using Mermaid syntax:
+
+      :::artifact{identifier="tea-making-flowchart" type="application/vnd.mermaid" title="Flow chart: Making Tea"}
+      \`\`\`mermaid
+      graph TD
+          A[Start] --> B{Water boiled?}
+          B -->|Yes| C[Add tea leaves to cup]
+          B -->|No| D[Boil water]
+          D --> B
+          C --> E[Pour boiling water into cup]
+          E --> F[Steep tea for desired time]
+          F --> G[Remove tea leaves]
+          G --> H[Add milk or sugar, if desired]
+          H --> I[Enjoy your tea!]
+          I --> J[End]
+      \`\`\`
+      :::
+
+      This flow chart uses Mermaid syntax to visualize the steps involved in making a cup of tea. Here's a brief explanation of the process:
+
+      1. Start
+      2. Check if water is boiled
+      3. If not boiled, boil the water
+      4. Once water is boiled, add tea leaves to the cup
+      5. Pour boiling water into the cup
+      6. Steep the tea for the desired time
+      7. Remove the tea leaves
+      8. Optionally add milk or sugar
+      9. Enjoy your tea!
+      10. End
+
+      This chart provides a clear visual representation of the tea-making process. You can easily modify or expand this chart if you want to add more details or steps to the process. Let me know if you'd like any changes or have any questions!
+    </assistant_response>
+  </example>
+
+  <example>
+    <user_query>Create a simple React counter component</user_query>
+    <assistant_response>
+      Here's a simple React counter component:
+
+      :::artifact{identifier="react-counter" type="application/vnd.react" title="React Counter"}
+      \`\`\`
+      import { useState } from 'react';
+
+      export default function Counter() {
+        const [count, setCount] = useState(0);
+        return (
+          <div className="p-4">
+            <p className="mb-2">Count: {count}</p>
+            <button className="bg-blue-500 text-white px-4 py-2 rounded" onClick={() => setCount(count + 1)}>
+              Increment
+            </button>
+          </div>
+        );
+      }
+      \`\`\`
+      :::
+
+      This component creates a simple counter with an increment button.
+    </assistant_response>
+  </example>
+
+  <example>
+    <user_query>Create a basic HTML structure for a blog post</user_query>
+    <assistant_response>
+      Here's a basic HTML structure for a blog post:
+
+      :::artifact{identifier="blog-post-html" type="text/html" title="Blog Post HTML"}
+      \`\`\`
+      <!DOCTYPE html>
+      <html lang="en">
+      <head>
+        <meta charset="UTF-8">
+        <meta name="viewport" content="width=device-width, initial-scale=1.0">
+        <title>My Blog Post</title>
+        <style>
+          body { font-family: Arial, sans-serif; line-height: 1.6; max-width: 800px; margin: 0 auto; padding: 20px; }
+          h1 { color: #333; }
+          p { margin-bottom: 15px; }
+        </style>
+      </head>
+      <body>
+        <header>
+          <h1>My First Blog Post</h1>
+        </header>
+        <main>
+          <article>
+            <p>This is the content of my blog post. It's short and sweet!</p>
+          </article>
+        </main>
+        <footer>
+          <p>&copy; 2023 My Blog</p>
+        </footer>
+      </body>
+      </html>
+      \`\`\`
+      :::
+
+      This HTML structure provides a simple layout for a blog post.
+    </assistant_response>
+  </example>
+</examples>`;
+
+const artifactsOpenAIPrompt = dedent`The assistant can create and reference artifacts during conversations.
+  
+Artifacts are for substantial, self-contained content that users might modify or reuse, displayed in a separate UI window for clarity.
+
+# Good artifacts are...
+- Substantial content (>15 lines)
+- Content that the user is likely to modify, iterate on, or take ownership of
+- Self-contained, complex content that can be understood on its own, without context from the conversation
+- Content intended for eventual use outside the conversation (e.g., reports, emails, presentations)
+- Content likely to be referenced or reused multiple times
+
+# Don't use artifacts for...
+- Simple, informational, or short content, such as brief code snippets, mathematical equations, or small examples
+- Primarily explanatory, instructional, or illustrative content, such as examples provided to clarify a concept
+- Suggestions, commentary, or feedback on existing artifacts
+- Conversational or explanatory content that doesn't represent a standalone piece of work
+- Content that is dependent on the current conversational context to be useful
+- Content that is unlikely to be modified or iterated upon by the user
+- Request from users that appears to be a one-off question
+
+# Usage notes
+- One artifact per message unless specifically requested
+- Prefer in-line content (don't use artifacts) when possible. Unnecessary use of artifacts can be jarring for users.
+- If a user asks the assistant to "draw an SVG" or "make a website," the assistant does not need to explain that it doesn't have these capabilities. Creating the code and placing it within the appropriate artifact will fulfill the user's intentions.
+- If asked to generate an image, the assistant can offer an SVG instead. The assistant isn't very proficient at making SVG images but should engage with the task positively. Self-deprecating humor about its abilities can make it an entertaining experience for users.
+- The assistant errs on the side of simplicity and avoids overusing artifacts for content that can be effectively presented within the conversation.
+- Always provide complete, specific, and fully functional content for artifacts without any snippets, placeholders, ellipses, or 'remains the same' comments.
+- If an artifact is not necessary or requested, the assistant should not mention artifacts at all, and respond to the user accordingly.
+
+## Artifact Instructions
+  When collaborating with the user on creating content that falls into compatible categories, the assistant should follow these steps:
+
+  1. Create the artifact using the following remark-directive markdown format:
+
+      :::artifact{identifier="unique-identifier" type="mime-type" title="Artifact Title"}
+      \`\`\`
+      Your artifact content here
+      \`\`\`
+      :::
+
+  a. Example of correct format:
+
+      :::artifact{identifier="example-artifact" type="text/plain" title="Example Artifact"}
+      \`\`\`
+      This is the content of the artifact.
+      It can span multiple lines.
+      \`\`\`
+      :::
+
+  b. Common mistakes to avoid:
+   - Don't split the opening ::: line
+   - Don't add extra backticks outside the artifact structure
+   - Don't omit the closing :::
+
+  2. Assign an identifier to the \`identifier\` attribute. For updates, reuse the prior identifier. For new artifacts, the identifier should be descriptive and relevant to the content, using kebab-case (e.g., "example-code-snippet"). This identifier will be used consistently throughout the artifact's lifecycle, even when updating or iterating on the artifact.
+  3. Include a \`title\` attribute to provide a brief title or description of the content.
+  4. Add a \`type\` attribute to specify the type of content the artifact represents. Assign one of the following values to the \`type\` attribute:
+    - HTML: "text/html"
+      - The user interface can render single file HTML pages placed within the artifact tags. HTML, JS, and CSS should be in a single file when using the \`text/html\` type.
+      - Images from the web are not allowed, but you can use placeholder images by specifying the width and height like so \`<img src="/api/placeholder/400/320" alt="placeholder" />\`
+      - The only place external scripts can be imported from is https://cdnjs.cloudflare.com
+    - SVG: "image/svg+xml"
+      - The user interface will render the Scalable Vector Graphics (SVG) image within the artifact tags.
+      - The assistant should specify the viewbox of the SVG rather than defining a width/height
+    - Mermaid Diagrams: "application/vnd.mermaid"
+      - The user interface will render Mermaid diagrams placed within the artifact tags.
+    - React Components: "application/vnd.react"
+      - Use this for displaying either: React elements, e.g. \`<strong>Hello World!</strong>\`, React pure functional components, e.g. \`() => <strong>Hello World!</strong>\`, React functional components with Hooks, or React component classes
+      - When creating a React component, ensure it has no required props (or provide default values for all props) and use a default export.
+      - Use Tailwind classes for styling. DO NOT USE ARBITRARY VALUES (e.g. \`h-[600px]\`).
+      - Base React is available to be imported. To use hooks, first import it at the top of the artifact, e.g. \`import { useState } from "react"\`
+      - The lucide-react@0.394.0 library is available to be imported. e.g. \`import { Camera } from "lucide-react"\` & \`<Camera color="red" size={48} />\`
+      - The recharts charting library is available to be imported, e.g. \`import { LineChart, XAxis, ... } from "recharts"\` & \`<LineChart ...><XAxis dataKey="name"> ...\`
+      - The three.js library is available to be imported, e.g. \`import * as THREE from "three";\`
+      - The date-fns library is available to be imported, e.g. \`import { compareAsc, format } from "date-fns";\`
+      - The react-day-picker library is available to be imported, e.g. \`import { DayPicker } from "react-day-picker";\`
+      - The assistant can use prebuilt components from the \`shadcn/ui\` library after it is imported: \`import { Alert, AlertDescription, AlertTitle, AlertDialog, AlertDialogAction } from '/components/ui/alert';\`. If using components from the shadcn/ui library, the assistant mentions this to the user and offers to help them install the components if necessary.
+      - Components MUST be imported from \`/components/ui/name\` and NOT from \`/components/name\` or \`@/components/ui/name\`.
+      - NO OTHER LIBRARIES (e.g. zod, hookform) ARE INSTALLED OR ABLE TO BE IMPORTED.
+      - Images from the web are not allowed, but you can use placeholder images by specifying the width and height like so \`<img src="/api/placeholder/400/320" alt="placeholder" />\`
+      - When iterating on code, ensure that the code is complete and functional without any snippets, placeholders, or ellipses.
+      - If you are unable to follow the above requirements for any reason, don't use artifacts and use regular code blocks instead, which will not attempt to render the component.
+  5. Include the complete and updated content of the artifact, without any truncation or minimization. Don't use "// rest of the code remains the same...".
+  6. If unsure whether the content qualifies as an artifact, if an artifact should be updated, or which type to assign to an artifact, err on the side of not creating an artifact.
+  7. NEVER use triple backticks to enclose the artifact, ONLY the content within the artifact.
+
+Here are some examples of correct usage of artifacts:
+
+## Examples
+
+### Example 1
+
+    This example demonstrates how to create a Mermaid artifact for a simple flow chart.
+
+    User: Can you create a simple flow chart showing the process of making tea using Mermaid?
+
+    Assistant: Sure! Here's a simple flow chart depicting the process of making tea using Mermaid syntax:
+
+      :::artifact{identifier="tea-making-flowchart" type="application/vnd.mermaid" title="Flow chart: Making Tea"}
+      \`\`\`mermaid
+      graph TD
+          A[Start] --> B{Water boiled?}
+          B -->|Yes| C[Add tea leaves to cup]
+          B -->|No| D[Boil water]
+          D --> B
+          C --> E[Pour boiling water into cup]
+          E --> F[Steep tea for desired time]
+          F --> G[Remove tea leaves]
+          G --> H[Add milk or sugar, if desired]
+          H --> I[Enjoy your tea!]
+          I --> J[End]
+      \`\`\`
+      :::
+
+      This flow chart uses Mermaid syntax to visualize the steps involved in making a cup of tea. Here's a brief explanation of the process:
+
+      1. Start
+      2. Check if water is boiled
+      3. If not boiled, boil the water
+      4. Once water is boiled, add tea leaves to the cup
+      5. Pour boiling water into the cup
+      6. Steep the tea for the desired time
+      7. Remove the tea leaves
+      8. Optionally add milk or sugar
+      9. Enjoy your tea!
+      10. End
+
+      This chart provides a clear visual representation of the tea-making process. You can easily modify or expand this chart if you want to add more details or steps to the process. Let me know if you'd like any changes or have any questions!
+
+---
+
+### Example 2
+
+    User: Create a simple React counter component
+    
+    Assistant: Here's a simple React counter component:
+
+      :::artifact{identifier="react-counter" type="application/vnd.react" title="React Counter"}
+      \`\`\`
+      import { useState } from 'react';
+
+      export default function Counter() {
+        const [count, setCount] = useState(0);
+        return (
+          <div className="p-4">
+            <p className="mb-2">Count: {count}</p>
+            <button className="bg-blue-500 text-white px-4 py-2 rounded" onClick={() => setCount(count + 1)}>
+              Increment
+            </button>
+          </div>
+        );
+      }
+      \`\`\`
+      :::
+
+      This component creates a simple counter with an increment button.
+
+---
+
+### Example 3
+    User: Create a basic HTML structure for a blog post
+    Assistant: Here's a basic HTML structure for a blog post:
+
+      :::artifact{identifier="blog-post-html" type="text/html" title="Blog Post HTML"}
+      \`\`\`
+      <!DOCTYPE html>
+      <html lang="en">
+      <head>
+        <meta charset="UTF-8">
+        <meta name="viewport" content="width=device-width, initial-scale=1.0">
+        <title>My Blog Post</title>
+        <style>
+          body { font-family: Arial, sans-serif; line-height: 1.6; max-width: 800px; margin: 0 auto; padding: 20px; }
+          h1 { color: #333; }
+          p { margin-bottom: 15px; }
+        </style>
+      </head>
+      <body>
+        <header>
+          <h1>My First Blog Post</h1>
+        </header>
+        <main>
+          <article>
+            <p>This is the content of my blog post. It's short and sweet!</p>
+          </article>
+        </main>
+        <footer>
+          <p>&copy; 2023 My Blog</p>
+        </footer>
+      </body>
+      </html>
+      \`\`\`
+      :::
+
+      This HTML structure provides a simple layout for a blog post.
+
+---`;
+
+/**
+ *
+ * @param {Object} params
+ * @param {EModelEndpoint | string} params.endpoint - The current endpoint
+ * @param {ArtifactModes} params.artifacts - The current artifact mode
+ * @returns
+ */
+const generateArtifactsPrompt = ({ endpoint, artifacts }) => {
+  if (artifacts === ArtifactModes.CUSTOM) {
+    return null;
+  }
+
+  let prompt = artifactsPrompt;
+  if (endpoint !== EModelEndpoint.anthropic) {
+    prompt = artifactsOpenAIPrompt;
+  }
+
+  if (artifacts === ArtifactModes.SHADCNUI) {
+    prompt += generateShadcnPrompt({ components, useXML: endpoint === EModelEndpoint.anthropic });
+  }
+
+  return prompt;
+};
+
+module.exports = generateArtifactsPrompt;

api/app/clients/prompts/createContextHandlers.js

@@ -0,0 +1,160 @@
+const axios = require('axios');
+const { isEnabled } = require('~/server/utils');
+const { logger } = require('~/config');
+
+const footer = `Use the context as your learned knowledge to better answer the user.
+
+In your response, remember to follow these guidelines:
+- If you don't know the answer, simply say that you don't know.
+- If you are unsure how to answer, ask for clarification.
+- Avoid mentioning that you obtained the information from the context.
+`;
+
+function createContextHandlers(req, userMessageContent) {
+  if (!process.env.RAG_API_URL) {
+    return;
+  }
+
+  const queryPromises = [];
+  const processedFiles = [];
+  const processedIds = new Set();
+  const jwtToken = req.headers.authorization.split(' ')[1];
+  const useFullContext = isEnabled(process.env.RAG_USE_FULL_CONTEXT);
+
+  const query = async (file) => {
+    if (useFullContext) {
+      return axios.get(`${process.env.RAG_API_URL}/documents/${file.file_id}/context`, {
+        headers: {
+          Authorization: `Bearer ${jwtToken}`,
+        },
+      });
+    }
+
+    return axios.post(
+      `${process.env.RAG_API_URL}/query`,
+      {
+        file_id: file.file_id,
+        query: userMessageContent,
+        k: 4,
+      },
+      {
+        headers: {
+          Authorization: `Bearer ${jwtToken}`,
+          'Content-Type': 'application/json',
+        },
+      },
+    );
+  };
+
+  const processFile = async (file) => {
+    if (file.embedded && !processedIds.has(file.file_id)) {
+      try {
+        const promise = query(file);
+        queryPromises.push(promise);
+        processedFiles.push(file);
+        processedIds.add(file.file_id);
+      } catch (error) {
+        logger.error(`Error processing file ${file.filename}:`, error);
+      }
+    }
+  };
+
+  const createContext = async () => {
+    try {
+      if (!queryPromises.length || !processedFiles.length) {
+        return '';
+      }
+
+      const oneFile = processedFiles.length === 1;
+      const header = `The user has attached ${oneFile ? 'a' : processedFiles.length} file${
+        !oneFile ? 's' : ''
+      } to the conversation:`;
+
+      const files = `${
+        oneFile
+          ? ''
+          : `
+      <files>`
+      }${processedFiles
+        .map(
+          (file) => `
+              <file>
+                <filename>${file.filename}</filename>
+                <type>${file.type}</type>
+              </file>`,
+        )
+        .join('')}${
+        oneFile
+          ? ''
+          : `
+        </files>`
+      }`;
+
+      const resolvedQueries = await Promise.all(queryPromises);
+
+      const context =
+        resolvedQueries.length === 0
+          ? '\n\tThe semantic search did not return any results.'
+          : resolvedQueries
+            .map((queryResult, index) => {
+              const file = processedFiles[index];
+              let contextItems = queryResult.data;
+
+              const generateContext = (currentContext) =>
+                `
+          <file>
+            <filename>${file.filename}</filename>
+            <context>${currentContext}
+            </context>
+          </file>`;
+
+              if (useFullContext) {
+                return generateContext(`\n${contextItems}`);
+              }
+
+              contextItems = queryResult.data
+                .map((item) => {
+                  const pageContent = item[0].page_content;
+                  return `
+            <contextItem>
+              <![CDATA[${pageContent?.trim()}]]>
+            </contextItem>`;
+                })
+                .join('');
+
+              return generateContext(contextItems);
+            })
+            .join('');
+
+      if (useFullContext) {
+        const prompt = `${header}
+          ${context}
+          ${footer}`;
+
+        return prompt;
+      }
+
+      const prompt = `${header}
+        ${files}
+
+        A semantic search was executed with the user's message as the query, retrieving the following context inside <context></context> XML tags.
+
+        <context>${context}
+        </context>
+
+        ${footer}`;
+
+      return prompt;
+    } catch (error) {
+      logger.error('Error creating context:', error);
+      throw error;
+    }
+  };
+
+  return {
+    processFile,
+    createContext,
+  };
+}
+
+module.exports = createContextHandlers;

api/app/clients/prompts/createVisionPrompt.js

@@ -0,0 +1,34 @@
+/**
+ * Generates a prompt instructing the user to describe an image in detail, tailored to different types of visual content.
+ * @param {boolean} pluralized - Whether to pluralize the prompt for multiple images.
+ * @returns {string} - The generated vision prompt.
+ */
+const createVisionPrompt = (pluralized = false) => {
+  return `Please describe the image${
+    pluralized ? 's' : ''
+  } in detail, covering relevant aspects such as:
+
+  For photographs, illustrations, or artwork:
+  - The main subject(s) and their appearance, positioning, and actions
+  - The setting, background, and any notable objects or elements
+  - Colors, lighting, and overall mood or atmosphere
+  - Any interesting details, textures, or patterns
+  - The style, technique, or medium used (if discernible)
+  
+  For screenshots or images containing text:
+  - The content and purpose of the text
+  - The layout, formatting, and organization of the information
+  - Any notable visual elements, such as logos, icons, or graphics
+  - The overall context or message conveyed by the screenshot
+  
+  For graphs, charts, or data visualizations:
+  - The type of graph or chart (e.g., bar graph, line chart, pie chart)
+  - The variables being compared or analyzed
+  - Any trends, patterns, or outliers in the data
+  - The axis labels, scales, and units of measurement
+  - The title, legend, and any additional context provided
+  
+  Be as specific and descriptive as possible while maintaining clarity and concision.`;
+};
+
+module.exports = createVisionPrompt;

api/app/clients/prompts/formatAgentMessages.spec.js

@@ -0,0 +1,328 @@
+const { ToolMessage } = require('@langchain/core/messages');
+const { ContentTypes } = require('librechat-data-provider');
+const { HumanMessage, AIMessage, SystemMessage } = require('@langchain/core/messages');
+const { formatAgentMessages } = require('./formatMessages');
+
+describe('formatAgentMessages', () => {
+  it('should format simple user and AI messages', () => {
+    const payload = [
+      { role: 'user', content: 'Hello' },
+      { role: 'assistant', content: 'Hi there!' },
+    ];
+    const result = formatAgentMessages(payload);
+    expect(result).toHaveLength(2);
+    expect(result[0]).toBeInstanceOf(HumanMessage);
+    expect(result[1]).toBeInstanceOf(AIMessage);
+  });
+
+  it('should handle system messages', () => {
+    const payload = [{ role: 'system', content: 'You are a helpful assistant.' }];
+    const result = formatAgentMessages(payload);
+    expect(result).toHaveLength(1);
+    expect(result[0]).toBeInstanceOf(SystemMessage);
+  });
+
+  it('should format messages with content arrays', () => {
+    const payload = [
+      {
+        role: 'user',
+        content: [{ type: ContentTypes.TEXT, [ContentTypes.TEXT]: 'Hello' }],
+      },
+    ];
+    const result = formatAgentMessages(payload);
+    expect(result).toHaveLength(1);
+    expect(result[0]).toBeInstanceOf(HumanMessage);
+  });
+
+  it('should handle tool calls and create ToolMessages', () => {
+    const payload = [
+      {
+        role: 'assistant',
+        content: [
+          {
+            type: ContentTypes.TEXT,
+            [ContentTypes.TEXT]: 'Let me check that for you.',
+            tool_call_ids: ['123'],
+          },
+          {
+            type: ContentTypes.TOOL_CALL,
+            tool_call: {
+              id: '123',
+              name: 'search',
+              args: '{"query":"weather"}',
+              output: 'The weather is sunny.',
+            },
+          },
+        ],
+      },
+    ];
+    const result = formatAgentMessages(payload);
+    expect(result).toHaveLength(2);
+    expect(result[0]).toBeInstanceOf(AIMessage);
+    expect(result[1]).toBeInstanceOf(ToolMessage);
+    expect(result[0].tool_calls).toHaveLength(1);
+    expect(result[1].tool_call_id).toBe('123');
+  });
+
+  it('should handle multiple content parts in assistant messages', () => {
+    const payload = [
+      {
+        role: 'assistant',
+        content: [
+          { type: ContentTypes.TEXT, [ContentTypes.TEXT]: 'Part 1' },
+          { type: ContentTypes.TEXT, [ContentTypes.TEXT]: 'Part 2' },
+        ],
+      },
+    ];
+    const result = formatAgentMessages(payload);
+    expect(result).toHaveLength(1);
+    expect(result[0]).toBeInstanceOf(AIMessage);
+    expect(result[0].content).toHaveLength(2);
+  });
+
+  it('should throw an error for invalid tool call structure', () => {
+    const payload = [
+      {
+        role: 'assistant',
+        content: [
+          {
+            type: ContentTypes.TOOL_CALL,
+            tool_call: {
+              id: '123',
+              name: 'search',
+              args: '{"query":"weather"}',
+              output: 'The weather is sunny.',
+            },
+          },
+        ],
+      },
+    ];
+    expect(() => formatAgentMessages(payload)).toThrow('Invalid tool call structure');
+  });
+
+  it('should handle tool calls with non-JSON args', () => {
+    const payload = [
+      {
+        role: 'assistant',
+        content: [
+          { type: ContentTypes.TEXT, [ContentTypes.TEXT]: 'Checking...', tool_call_ids: ['123'] },
+          {
+            type: ContentTypes.TOOL_CALL,
+            tool_call: {
+              id: '123',
+              name: 'search',
+              args: 'non-json-string',
+              output: 'Result',
+            },
+          },
+        ],
+      },
+    ];
+    const result = formatAgentMessages(payload);
+    expect(result).toHaveLength(2);
+    expect(result[0].tool_calls[0].args).toStrictEqual({ input: 'non-json-string' });
+  });
+
+  it('should handle complex tool calls with multiple steps', () => {
+    const payload = [
+      {
+        role: 'assistant',
+        content: [
+          {
+            type: ContentTypes.TEXT,
+            [ContentTypes.TEXT]: 'I\'ll search for that information.',
+            tool_call_ids: ['search_1'],
+          },
+          {
+            type: ContentTypes.TOOL_CALL,
+            tool_call: {
+              id: 'search_1',
+              name: 'search',
+              args: '{"query":"weather in New York"}',
+              output: 'The weather in New York is currently sunny with a temperature of 75°F.',
+            },
+          },
+          {
+            type: ContentTypes.TEXT,
+            [ContentTypes.TEXT]: 'Now, I\'ll convert the temperature.',
+            tool_call_ids: ['convert_1'],
+          },
+          {
+            type: ContentTypes.TOOL_CALL,
+            tool_call: {
+              id: 'convert_1',
+              name: 'convert_temperature',
+              args: '{"temperature": 75, "from": "F", "to": "C"}',
+              output: '23.89°C',
+            },
+          },
+          { type: ContentTypes.TEXT, [ContentTypes.TEXT]: 'Here\'s your answer.' },
+        ],
+      },
+    ];
+
+    const result = formatAgentMessages(payload);
+
+    expect(result).toHaveLength(5);
+    expect(result[0]).toBeInstanceOf(AIMessage);
+    expect(result[1]).toBeInstanceOf(ToolMessage);
+    expect(result[2]).toBeInstanceOf(AIMessage);
+    expect(result[3]).toBeInstanceOf(ToolMessage);
+    expect(result[4]).toBeInstanceOf(AIMessage);
+
+    // Check first AIMessage
+    expect(result[0].content).toBe('I\'ll search for that information.');
+    expect(result[0].tool_calls).toHaveLength(1);
+    expect(result[0].tool_calls[0]).toEqual({
+      id: 'search_1',
+      name: 'search',
+      args: { query: 'weather in New York' },
+    });
+
+    // Check first ToolMessage
+    expect(result[1].tool_call_id).toBe('search_1');
+    expect(result[1].name).toBe('search');
+    expect(result[1].content).toBe(
+      'The weather in New York is currently sunny with a temperature of 75°F.',
+    );
+
+    // Check second AIMessage
+    expect(result[2].content).toBe('Now, I\'ll convert the temperature.');
+    expect(result[2].tool_calls).toHaveLength(1);
+    expect(result[2].tool_calls[0]).toEqual({
+      id: 'convert_1',
+      name: 'convert_temperature',
+      args: { temperature: 75, from: 'F', to: 'C' },
+    });
+
+    // Check second ToolMessage
+    expect(result[3].tool_call_id).toBe('convert_1');
+    expect(result[3].name).toBe('convert_temperature');
+    expect(result[3].content).toBe('23.89°C');
+
+    // Check final AIMessage
+    expect(result[4].content).toStrictEqual([
+      { [ContentTypes.TEXT]: 'Here\'s your answer.', type: ContentTypes.TEXT },
+    ]);
+  });
+
+  it.skip('should not produce two consecutive assistant messages and format content correctly', () => {
+    const payload = [
+      { role: 'user', content: 'Hello' },
+      {
+        role: 'assistant',
+        content: [{ type: ContentTypes.TEXT, [ContentTypes.TEXT]: 'Hi there!' }],
+      },
+      {
+        role: 'assistant',
+        content: [{ type: ContentTypes.TEXT, [ContentTypes.TEXT]: 'How can I help you?' }],
+      },
+      { role: 'user', content: 'What\'s the weather?' },
+      {
+        role: 'assistant',
+        content: [
+          {
+            type: ContentTypes.TEXT,
+            [ContentTypes.TEXT]: 'Let me check that for you.',
+            tool_call_ids: ['weather_1'],
+          },
+          {
+            type: ContentTypes.TOOL_CALL,
+            tool_call: {
+              id: 'weather_1',
+              name: 'check_weather',
+              args: '{"location":"New York"}',
+              output: 'Sunny, 75°F',
+            },
+          },
+        ],
+      },
+      {
+        role: 'assistant',
+        content: [
+          { type: ContentTypes.TEXT, [ContentTypes.TEXT]: 'Here\'s the weather information.' },
+        ],
+      },
+    ];
+
+    const result = formatAgentMessages(payload);
+
+    // Check correct message count and types
+    expect(result).toHaveLength(6);
+    expect(result[0]).toBeInstanceOf(HumanMessage);
+    expect(result[1]).toBeInstanceOf(AIMessage);
+    expect(result[2]).toBeInstanceOf(HumanMessage);
+    expect(result[3]).toBeInstanceOf(AIMessage);
+    expect(result[4]).toBeInstanceOf(ToolMessage);
+    expect(result[5]).toBeInstanceOf(AIMessage);
+
+    // Check content of messages
+    expect(result[0].content).toStrictEqual([
+      { [ContentTypes.TEXT]: 'Hello', type: ContentTypes.TEXT },
+    ]);
+    expect(result[1].content).toStrictEqual([
+      { [ContentTypes.TEXT]: 'Hi there!', type: ContentTypes.TEXT },
+      { [ContentTypes.TEXT]: 'How can I help you?', type: ContentTypes.TEXT },
+    ]);
+    expect(result[2].content).toStrictEqual([
+      { [ContentTypes.TEXT]: 'What\'s the weather?', type: ContentTypes.TEXT },
+    ]);
+    expect(result[3].content).toBe('Let me check that for you.');
+    expect(result[4].content).toBe('Sunny, 75°F');
+    expect(result[5].content).toStrictEqual([
+      { [ContentTypes.TEXT]: 'Here\'s the weather information.', type: ContentTypes.TEXT },
+    ]);
+
+    // Check that there are no consecutive AIMessages
+    const messageTypes = result.map((message) => message.constructor);
+    for (let i = 0; i < messageTypes.length - 1; i++) {
+      expect(messageTypes[i] === AIMessage && messageTypes[i + 1] === AIMessage).toBe(false);
+    }
+
+    // Additional check to ensure the consecutive assistant messages were combined
+    expect(result[1].content).toHaveLength(2);
+  });
+
+  it('should skip THINK type content parts', () => {
+    const payload = [
+      {
+        role: 'assistant',
+        content: [
+          { type: ContentTypes.TEXT, [ContentTypes.TEXT]: 'Initial response' },
+          { type: ContentTypes.THINK, [ContentTypes.THINK]: 'Reasoning about the problem...' },
+          { type: ContentTypes.TEXT, [ContentTypes.TEXT]: 'Final answer' },
+        ],
+      },
+    ];
+
+    const result = formatAgentMessages(payload);
+
+    expect(result).toHaveLength(1);
+    expect(result[0]).toBeInstanceOf(AIMessage);
+    expect(result[0].content).toEqual('Initial response\nFinal answer');
+  });
+
+  it('should join TEXT content as string when THINK content type is present', () => {
+    const payload = [
+      {
+        role: 'assistant',
+        content: [
+          { type: ContentTypes.THINK, [ContentTypes.THINK]: 'Analyzing the problem...' },
+          { type: ContentTypes.TEXT, [ContentTypes.TEXT]: 'First part of response' },
+          { type: ContentTypes.TEXT, [ContentTypes.TEXT]: 'Second part of response' },
+          { type: ContentTypes.TEXT, [ContentTypes.TEXT]: 'Final part of response' },
+        ],
+      },
+    ];
+
+    const result = formatAgentMessages(payload);
+
+    expect(result).toHaveLength(1);
+    expect(result[0]).toBeInstanceOf(AIMessage);
+    expect(typeof result[0].content).toBe('string');
+    expect(result[0].content).toBe(
+      'First part of response\nSecond part of response\nFinal part of response',
+    );
+    expect(result[0].content).not.toContain('Analyzing the problem...');
+  });
+});

api/app/clients/prompts/formatGoogleInputs.js

@@ -0,0 +1,42 @@
+/**
+ * Formats an object to match the struct_val, list_val, string_val, float_val, and int_val format.
+ *
+ * @param {Object} obj - The object to be formatted.
+ * @returns {Object} The formatted object.
+ *
+ * Handles different types:
+ * - Arrays are wrapped in list_val and each element is processed.
+ * - Objects are recursively processed.
+ * - Strings are wrapped in string_val.
+ * - Numbers are wrapped in float_val or int_val depending on whether they are floating-point or integers.
+ */
+function formatGoogleInputs(obj) {
+  const formattedObj = {};
+
+  for (const key in obj) {
+    if (Object.prototype.hasOwnProperty.call(obj, key)) {
+      const value = obj[key];
+
+      // Handle arrays
+      if (Array.isArray(value)) {
+        formattedObj[key] = { list_val: value.map((item) => formatGoogleInputs(item)) };
+      }
+      // Handle objects
+      else if (typeof value === 'object' && value !== null) {
+        formattedObj[key] = formatGoogleInputs(value);
+      }
+      // Handle numbers
+      else if (typeof value === 'number') {
+        formattedObj[key] = Number.isInteger(value) ? { int_val: value } : { float_val: value };
+      }
+      // Handle other types (e.g., strings)
+      else {
+        formattedObj[key] = { string_val: [value] };
+      }
+    }
+  }
+
+  return { struct_val: formattedObj };
+}
+
+module.exports = formatGoogleInputs;

api/app/clients/prompts/formatGoogleInputs.spec.js

@@ -0,0 +1,274 @@
+const formatGoogleInputs = require('./formatGoogleInputs');
+
+describe('formatGoogleInputs', () => {
+  it('formats message correctly', () => {
+    const input = {
+      messages: [
+        {
+          content: 'hi',
+          author: 'user',
+        },
+      ],
+      context: 'context',
+      examples: [
+        {
+          input: {
+            author: 'user',
+            content: 'user input',
+          },
+          output: {
+            author: 'bot',
+            content: 'bot output',
+          },
+        },
+      ],
+      parameters: {
+        temperature: 0.2,
+        topP: 0.8,
+        topK: 40,
+        maxOutputTokens: 1024,
+      },
+    };
+
+    const expectedOutput = {
+      struct_val: {
+        messages: {
+          list_val: [
+            {
+              struct_val: {
+                content: {
+                  string_val: ['hi'],
+                },
+                author: {
+                  string_val: ['user'],
+                },
+              },
+            },
+          ],
+        },
+        context: {
+          string_val: ['context'],
+        },
+        examples: {
+          list_val: [
+            {
+              struct_val: {
+                input: {
+                  struct_val: {
+                    author: {
+                      string_val: ['user'],
+                    },
+                    content: {
+                      string_val: ['user input'],
+                    },
+                  },
+                },
+                output: {
+                  struct_val: {
+                    author: {
+                      string_val: ['bot'],
+                    },
+                    content: {
+                      string_val: ['bot output'],
+                    },
+                  },
+                },
+              },
+            },
+          ],
+        },
+        parameters: {
+          struct_val: {
+            temperature: {
+              float_val: 0.2,
+            },
+            topP: {
+              float_val: 0.8,
+            },
+            topK: {
+              int_val: 40,
+            },
+            maxOutputTokens: {
+              int_val: 1024,
+            },
+          },
+        },
+      },
+    };
+
+    const result = formatGoogleInputs(input);
+    expect(JSON.stringify(result)).toEqual(JSON.stringify(expectedOutput));
+  });
+
+  it('formats real payload parts', () => {
+    const input = {
+      instances: [
+        {
+          context: 'context',
+          examples: [
+            {
+              input: {
+                author: 'user',
+                content: 'user input',
+              },
+              output: {
+                author: 'bot',
+                content: 'user output',
+              },
+            },
+          ],
+          messages: [
+            {
+              author: 'user',
+              content: 'hi',
+            },
+          ],
+        },
+      ],
+      parameters: {
+        candidateCount: 1,
+        maxOutputTokens: 1024,
+        temperature: 0.2,
+        topP: 0.8,
+        topK: 40,
+      },
+    };
+    const expectedOutput = {
+      struct_val: {
+        instances: {
+          list_val: [
+            {
+              struct_val: {
+                context: { string_val: ['context'] },
+                examples: {
+                  list_val: [
+                    {
+                      struct_val: {
+                        input: {
+                          struct_val: {
+                            author: { string_val: ['user'] },
+                            content: { string_val: ['user input'] },
+                          },
+                        },
+                        output: {
+                          struct_val: {
+                            author: { string_val: ['bot'] },
+                            content: { string_val: ['user output'] },
+                          },
+                        },
+                      },
+                    },
+                  ],
+                },
+                messages: {
+                  list_val: [
+                    {
+                      struct_val: {
+                        author: { string_val: ['user'] },
+                        content: { string_val: ['hi'] },
+                      },
+                    },
+                  ],
+                },
+              },
+            },
+          ],
+        },
+        parameters: {
+          struct_val: {
+            candidateCount: { int_val: 1 },
+            maxOutputTokens: { int_val: 1024 },
+            temperature: { float_val: 0.2 },
+            topP: { float_val: 0.8 },
+            topK: { int_val: 40 },
+          },
+        },
+      },
+    };
+
+    const result = formatGoogleInputs(input);
+    expect(JSON.stringify(result)).toEqual(JSON.stringify(expectedOutput));
+  });
+
+  it('helps create valid payload parts', () => {
+    const instances = {
+      context: 'context',
+      examples: [
+        {
+          input: {
+            author: 'user',
+            content: 'user input',
+          },
+          output: {
+            author: 'bot',
+            content: 'user output',
+          },
+        },
+      ],
+      messages: [
+        {
+          author: 'user',
+          content: 'hi',
+        },
+      ],
+    };
+
+    const expectedInstances = {
+      struct_val: {
+        context: { string_val: ['context'] },
+        examples: {
+          list_val: [
+            {
+              struct_val: {
+                input: {
+                  struct_val: {
+                    author: { string_val: ['user'] },
+                    content: { string_val: ['user input'] },
+                  },
+                },
+                output: {
+                  struct_val: {
+                    author: { string_val: ['bot'] },
+                    content: { string_val: ['user output'] },
+                  },
+                },
+              },
+            },
+          ],
+        },
+        messages: {
+          list_val: [
+            {
+              struct_val: {
+                author: { string_val: ['user'] },
+                content: { string_val: ['hi'] },
+              },
+            },
+          ],
+        },
+      },
+    };
+
+    const parameters = {
+      candidateCount: 1,
+      maxOutputTokens: 1024,
+      temperature: 0.2,
+      topP: 0.8,
+      topK: 40,
+    };
+    const expectedParameters = {
+      struct_val: {
+        candidateCount: { int_val: 1 },
+        maxOutputTokens: { int_val: 1024 },
+        temperature: { float_val: 0.2 },
+        topP: { float_val: 0.8 },
+        topK: { int_val: 40 },
+      },
+    };
+
+    const instancesResult = formatGoogleInputs(instances);
+    const parametersResult = formatGoogleInputs(parameters);
+    expect(JSON.stringify(instancesResult)).toEqual(JSON.stringify(expectedInstances));
+    expect(JSON.stringify(parametersResult)).toEqual(JSON.stringify(expectedParameters));
+  });
+});

api/app/clients/prompts/formatMessages.js

@@ -1,4 +1,28 @@
-const { HumanMessage, AIMessage, SystemMessage } = require('langchain/schema');
+const { ToolMessage } = require('@langchain/core/messages');
+const { EModelEndpoint, ContentTypes } = require('librechat-data-provider');
+const { HumanMessage, AIMessage, SystemMessage } = require('@langchain/core/messages');
+
+/**
+ * Formats a message to OpenAI Vision API payload format.
+ *
+ * @param {Object} params - The parameters for formatting.
+ * @param {Object} params.message - The message object to format.
+ * @param {string} [params.message.role] - The role of the message sender (must be 'user').
+ * @param {string} [params.message.content] - The text content of the message.
+ * @param {EModelEndpoint} [params.endpoint] - Identifier for specific endpoint handling
+ * @param {Array<string>} [params.image_urls] - The image_urls to attach to the message.
+ * @returns {(Object)} - The formatted message.
+ */
+const formatVisionMessage = ({ message, image_urls, endpoint }) => {
+  if (endpoint === EModelEndpoint.anthropic) {
+    message.content = [...image_urls, { type: ContentTypes.TEXT, text: message.content }];
+    return message;
+  }
+
+  message.content = [{ type: ContentTypes.TEXT, text: message.content }, ...image_urls];
+
+  return message;
+};
 
 /**
  * Formats a message to OpenAI payload format based on the provided options.
@@ -10,12 +34,14 @@ const { HumanMessage, AIMessage, SystemMessage } = require('langchain/schema');
  * @param {string} [params.message.sender] - The sender of the message.
  * @param {string} [params.message.text] - The text content of the message.
  * @param {string} [params.message.content] - The content of the message.
+ * @param {Array<string>} [params.message.image_urls] - The image_urls attached to the message for Vision API.
  * @param {string} [params.userName] - The name of the user.
  * @param {string} [params.assistantName] - The name of the assistant.
+ * @param {string} [params.endpoint] - Identifier for specific endpoint handling
  * @param {boolean} [params.langChain=false] - Whether to return a LangChain message object.
  * @returns {(Object|HumanMessage|AIMessage|SystemMessage)} - The formatted message.
  */
-const formatMessage = ({ message, userName, assistantName, langChain = false }) => {
+const formatMessage = ({ message, userName, assistantName, endpoint, langChain = false }) => {
   let { role: _role, _name, sender, text, content: _content, lc_id } = message;
   if (lc_id && lc_id[2] && !langChain) {
     const roleMapping = {
@@ -26,12 +52,21 @@ const formatMessage = ({ message, userName, assistantName, langChain = false })
     _role = roleMapping[lc_id[2]];
   }
   const role = _role ?? (sender && sender?.toLowerCase() === 'user' ? 'user' : 'assistant');
-  const content = text ?? _content ?? '';
+  const content = _content ?? text ?? '';
   const formattedMessage = {
     role,
     content,
   };
 
+  const { image_urls } = message;
+  if (Array.isArray(image_urls) && image_urls.length > 0 && role === 'user') {
+    return formatVisionMessage({
+      message: formattedMessage,
+      image_urls: message.image_urls,
+      endpoint,
+    });
+  }
+
   if (_name) {
     formattedMessage.name = _name;
   }
@@ -97,4 +132,144 @@ const formatFromLangChain = (message) => {
   };
 };
 
-module.exports = { formatMessage, formatLangChainMessages, formatFromLangChain };
+/**
+ * Formats an array of messages for LangChain, handling tool calls and creating ToolMessage instances.
+ *
+ * @param {Array<Partial<TMessage>>} payload - The array of messages to format.
+ * @returns {Array<(HumanMessage|AIMessage|SystemMessage|ToolMessage)>} - The array of formatted LangChain messages, including ToolMessages for tool calls.
+ */
+const formatAgentMessages = (payload) => {
+  const messages = [];
+
+  for (const message of payload) {
+    if (typeof message.content === 'string') {
+      message.content = [{ type: ContentTypes.TEXT, [ContentTypes.TEXT]: message.content }];
+    }
+    if (message.role !== 'assistant') {
+      messages.push(formatMessage({ message, langChain: true }));
+      continue;
+    }
+
+    let currentContent = [];
+    let lastAIMessage = null;
+
+    let hasReasoning = false;
+    for (const part of message.content) {
+      if (part.type === ContentTypes.TEXT && part.tool_call_ids) {
+        /*
+        If there's pending content, it needs to be aggregated as a single string to prepare for tool calls.
+        For Anthropic models, the "tool_calls" field on a message is only respected if content is a string.
+         */
+        if (currentContent.length > 0) {
+          let content = currentContent.reduce((acc, curr) => {
+            if (curr.type === ContentTypes.TEXT) {
+              return `${acc}${curr[ContentTypes.TEXT]}\n`;
+            }
+            return acc;
+          }, '');
+          content = `${content}\n${part[ContentTypes.TEXT] ?? ''}`.trim();
+          lastAIMessage = new AIMessage({ content });
+          messages.push(lastAIMessage);
+          currentContent = [];
+          continue;
+        }
+
+        // Create a new AIMessage with this text and prepare for tool calls
+        lastAIMessage = new AIMessage({
+          content: part.text || '',
+        });
+
+        messages.push(lastAIMessage);
+      } else if (part.type === ContentTypes.TOOL_CALL) {
+        if (!lastAIMessage) {
+          throw new Error('Invalid tool call structure: No preceding AIMessage with tool_call_ids');
+        }
+
+        // Note: `tool_calls` list is defined when constructed by `AIMessage` class, and outputs should be excluded from it
+        const { output, args: _args, ...tool_call } = part.tool_call;
+        // TODO: investigate; args as dictionary may need to be provider-or-tool-specific
+        let args = _args;
+        try {
+          args = JSON.parse(_args);
+        } catch (e) {
+          if (typeof _args === 'string') {
+            args = { input: _args };
+          }
+        }
+
+        tool_call.args = args;
+        lastAIMessage.tool_calls.push(tool_call);
+
+        // Add the corresponding ToolMessage
+        messages.push(
+          new ToolMessage({
+            tool_call_id: tool_call.id,
+            name: tool_call.name,
+            content: output || '',
+          }),
+        );
+      } else if (part.type === ContentTypes.THINK) {
+        hasReasoning = true;
+        continue;
+      } else {
+        currentContent.push(part);
+      }
+    }
+
+    if (hasReasoning) {
+      currentContent = currentContent
+        .reduce((acc, curr) => {
+          if (curr.type === ContentTypes.TEXT) {
+            return `${acc}${curr[ContentTypes.TEXT]}\n`;
+          }
+          return acc;
+        }, '')
+        .trim();
+    }
+
+    if (currentContent.length > 0) {
+      messages.push(new AIMessage({ content: currentContent }));
+    }
+  }
+
+  return messages;
+};
+
+/**
+ * Formats an array of messages for LangChain, making sure all content fields are strings
+ * @param {Array<(HumanMessage|AIMessage|SystemMessage|ToolMessage)>} payload - The array of messages to format.
+ * @returns {Array<(HumanMessage|AIMessage|SystemMessage|ToolMessage)>} - The array of formatted LangChain messages, including ToolMessages for tool calls.
+ */
+const formatContentStrings = (payload) => {
+  const messages = [];
+
+  for (const message of payload) {
+    if (typeof message.content === 'string') {
+      continue;
+    }
+
+    if (!Array.isArray(message.content)) {
+      continue;
+    }
+
+    // Reduce text types to a single string, ignore all other types
+    const content = message.content.reduce((acc, curr) => {
+      if (curr.type === ContentTypes.TEXT) {
+        return `${acc}${curr[ContentTypes.TEXT]}\n`;
+      }
+      return acc;
+    }, '');
+
+    message.content = content.trim();
+  }
+
+  return messages;
+};
+
+module.exports = {
+  formatMessage,
+  formatFromLangChain,
+  formatAgentMessages,
+  formatContentStrings,
+  formatLangChainMessages,
+};

api/app/clients/prompts/formatMessages.spec.js

@@ -1,5 +1,6 @@
+const { Constants } = require('librechat-data-provider');
+const { HumanMessage, AIMessage, SystemMessage } = require('@langchain/core/messages');
 const { formatMessage, formatLangChainMessages, formatFromLangChain } = require('./formatMessages');
-const { HumanMessage, AIMessage, SystemMessage } = require('langchain/schema');
 
 describe('formatMessage', () => {
   it('formats user message', () => {
@@ -54,15 +55,13 @@ describe('formatMessage', () => {
         _id: '6512cdfb92cbf69fea615331',
         messageId: 'b620bf73-c5c3-4a38-b724-76886aac24c4',
         __v: 0,
-        cancelled: false,
         conversationId: '5c23d24f-941f-4aab-85df-127b596c8aa5',
         createdAt: Date.now(),
         error: false,
         finish_reason: null,
         isCreatedByUser: true,
-        isEdited: false,
         model: null,
-        parentMessageId: '00000000-0000-0000-0000-000000000000',
+        parentMessageId: Constants.NO_PARENT,
         sender: 'User',
         text: 'hi',
         tokenCount: 5,

api/app/clients/prompts/index.js

@@ -1,15 +1,21 @@
+const addCacheControl = require('./addCacheControl');
 const formatMessages = require('./formatMessages');
 const summaryPrompts = require('./summaryPrompts');
 const handleInputs = require('./handleInputs');
 const instructions = require('./instructions');
 const titlePrompts = require('./titlePrompts');
-const truncateText = require('./truncateText');
+const truncate = require('./truncate');
+const createVisionPrompt = require('./createVisionPrompt');
+const createContextHandlers = require('./createContextHandlers');
 
 module.exports = {
+  addCacheControl,
   ...formatMessages,
   ...summaryPrompts,
   ...handleInputs,
   ...instructions,
   ...titlePrompts,
-  truncateText,
+  ...truncate,
+  createVisionPrompt,
+  createContextHandlers,
 };

api/app/clients/prompts/shadcn-docs/components.js

@@ -0,0 +1,495 @@
+// Essential Components
+const essentialComponents = {
+  avatar: {
+    componentName: 'Avatar',
+    importDocs: 'import { Avatar, AvatarFallback, AvatarImage } from "/components/ui/avatar"',
+    usageDocs: `
+<Avatar>
+  <AvatarImage src="https://github.com/shadcn.png" />
+  <AvatarFallback>CN</AvatarFallback>
+</Avatar>`,
+  },
+  button: {
+    componentName: 'Button',
+    importDocs: 'import { Button } from "/components/ui/button"',
+    usageDocs: `
+<Button variant="outline">Button</Button>`,
+  },
+  card: {
+    componentName: 'Card',
+    importDocs: `
+import {
+  Card,
+  CardContent,
+  CardDescription,
+  CardFooter,
+  CardHeader,
+  CardTitle,
+} from "/components/ui/card"`,
+    usageDocs: `
+<Card>
+  <CardHeader>
+    <CardTitle>Card Title</CardTitle>
+    <CardDescription>Card Description</CardDescription>
+  </CardHeader>
+  <CardContent>
+    <p>Card Content</p>
+  </CardContent>
+  <CardFooter>
+    <p>Card Footer</p>
+  </CardFooter>
+</Card>`,
+  },
+  checkbox: {
+    componentName: 'Checkbox',
+    importDocs: 'import { Checkbox } from "/components/ui/checkbox"',
+    usageDocs: '<Checkbox />',
+  },
+  input: {
+    componentName: 'Input',
+    importDocs: 'import { Input } from "/components/ui/input"',
+    usageDocs: '<Input />',
+  },
+  label: {
+    componentName: 'Label',
+    importDocs: 'import { Label } from "/components/ui/label"',
+    usageDocs: '<Label htmlFor="email">Your email address</Label>',
+  },
+  radioGroup: {
+    componentName: 'RadioGroup',
+    importDocs: `
+import { Label } from "/components/ui/label"
+import { RadioGroup, RadioGroupItem } from "/components/ui/radio-group"`,
+    usageDocs: `
+<RadioGroup defaultValue="option-one">
+  <div className="flex items-center space-x-2">
+    <RadioGroupItem value="option-one" id="option-one" />
+    <Label htmlFor="option-one">Option One</Label>
+  </div>
+  <div className="flex items-center space-x-2">
+    <RadioGroupItem value="option-two" id="option-two" />
+    <Label htmlFor="option-two">Option Two</Label>
+  </div>
+</RadioGroup>`,
+  },
+  select: {
+    componentName: 'Select',
+    importDocs: `
+import {
+  Select,
+  SelectContent,
+  SelectItem,
+  SelectTrigger,
+  SelectValue,
+} from "/components/ui/select"`,
+    usageDocs: `
+<Select>
+  <SelectTrigger className="w-[180px]">
+    <SelectValue placeholder="Theme" />
+  </SelectTrigger>
+  <SelectContent>
+    <SelectItem value="light">Light</SelectItem>
+    <SelectItem value="dark">Dark</SelectItem>
+    <SelectItem value="system">System</SelectItem>
+  </SelectContent>
+</Select>`,
+  },
+  textarea: {
+    componentName: 'Textarea',
+    importDocs: 'import { Textarea } from "/components/ui/textarea"',
+    usageDocs: '<Textarea />',
+  },
+};
+
+// Extra Components
+const extraComponents = {
+  accordion: {
+    componentName: 'Accordion',
+    importDocs: `
+import {
+  Accordion,
+  AccordionContent,
+  AccordionItem,
+  AccordionTrigger,
+} from "/components/ui/accordion"`,
+    usageDocs: `
+<Accordion type="single" collapsible>
+  <AccordionItem value="item-1">
+    <AccordionTrigger>Is it accessible?</AccordionTrigger>
+    <AccordionContent>
+      Yes. It adheres to the WAI-ARIA design pattern.
+    </AccordionContent>
+  </AccordionItem>
+</Accordion>`,
+  },
+  alertDialog: {
+    componentName: 'AlertDialog',
+    importDocs: `
+import {
+  AlertDialog,
+  AlertDialogAction,
+  AlertDialogCancel,
+  AlertDialogContent,
+  AlertDialogDescription,
+  AlertDialogFooter,
+  AlertDialogHeader,
+  AlertDialogTitle,
+  AlertDialogTrigger,
+} from "/components/ui/alert-dialog"`,
+    usageDocs: `
+<AlertDialog>
+  <AlertDialogTrigger>Open</AlertDialogTrigger>
+  <AlertDialogContent>
+    <AlertDialogHeader>
+      <AlertDialogTitle>Are you absolutely sure?</AlertDialogTitle>
+      <AlertDialogDescription>
+        This action cannot be undone.
+      </AlertDialogDescription>
+    </AlertDialogHeader>
+    <AlertDialogFooter>
+      <AlertDialogCancel>Cancel</AlertDialogCancel>
+      <AlertDialogAction>Continue</AlertDialogAction>
+    </AlertDialogFooter>
+  </AlertDialogContent>
+</AlertDialog>`,
+  },
+  alert: {
+    componentName: 'Alert',
+    importDocs: `
+import {
+  Alert,
+  AlertDescription,
+  AlertTitle,
+} from "/components/ui/alert"`,
+    usageDocs: `
+<Alert>
+  <AlertTitle>Heads up!</AlertTitle>
+  <AlertDescription>
+    You can add components to your app using the cli.
+  </AlertDescription>
+</Alert>`,
+  },
+  aspectRatio: {
+    componentName: 'AspectRatio',
+    importDocs: 'import { AspectRatio } from "/components/ui/aspect-ratio"',
+    usageDocs: `
+<AspectRatio ratio={16 / 9}>
+  <Image src="..." alt="Image" className="rounded-md object-cover" />
+</AspectRatio>`,
+  },
+  badge: {
+    componentName: 'Badge',
+    importDocs: 'import { Badge } from "/components/ui/badge"',
+    usageDocs: '<Badge>Badge</Badge>',
+  },
+  calendar: {
+    componentName: 'Calendar',
+    importDocs: 'import { Calendar } from "/components/ui/calendar"',
+    usageDocs: '<Calendar />',
+  },
+  carousel: {
+    componentName: 'Carousel',
+    importDocs: `
+import {
+  Carousel,
+  CarouselContent,
+  CarouselItem,
+  CarouselNext,
+  CarouselPrevious,
+} from "/components/ui/carousel"`,
+    usageDocs: `
+<Carousel>
+  <CarouselContent>
+    <CarouselItem>...</CarouselItem>
+    <CarouselItem>...</CarouselItem>
+    <CarouselItem>...</CarouselItem>
+  </CarouselContent>
+  <CarouselPrevious />
+  <CarouselNext />
+</Carousel>`,
+  },
+  collapsible: {
+    componentName: 'Collapsible',
+    importDocs: `
+import {
+  Collapsible,
+  CollapsibleContent,
+  CollapsibleTrigger,
+} from "/components/ui/collapsible"`,
+    usageDocs: `
+<Collapsible>
+  <CollapsibleTrigger>Can I use this in my project?</CollapsibleTrigger>
+  <CollapsibleContent>
+    Yes. Free to use for personal and commercial projects. No attribution required.
+  </CollapsibleContent>
+</Collapsible>`,
+  },
+  dialog: {
+    componentName: 'Dialog',
+    importDocs: `
+import {
+  Dialog,
+  DialogContent,
+  DialogDescription,
+  DialogHeader,
+  DialogTitle,
+  DialogTrigger,
+} from "/components/ui/dialog"`,
+    usageDocs: `
+<Dialog>
+  <DialogTrigger>Open</DialogTrigger>
+  <DialogContent>
+    <DialogHeader>
+      <DialogTitle>Are you sure absolutely sure?</DialogTitle>
+      <DialogDescription>
+        This action cannot be undone.
+      </DialogDescription>
+    </DialogHeader>
+  </DialogContent>
+</Dialog>`,
+  },
+  dropdownMenu: {
+    componentName: 'DropdownMenu',
+    importDocs: `
+import {
+  DropdownMenu,
+  DropdownMenuContent,
+  DropdownMenuItem,
+  DropdownMenuLabel,
+  DropdownMenuSeparator,
+  DropdownMenuTrigger,
+} from "/components/ui/dropdown-menu"`,
+    usageDocs: `
+<DropdownMenu>
+  <DropdownMenuTrigger>Open</DropdownMenuTrigger>
+  <DropdownMenuContent>
+    <DropdownMenuLabel>My Account</DropdownMenuLabel>
+    <DropdownMenuSeparator />
+    <DropdownMenuItem>Profile</DropdownMenuItem>
+    <DropdownMenuItem>Billing</DropdownMenuItem>
+    <DropdownMenuItem>Team</DropdownMenuItem>
+    <DropdownMenuItem>Subscription</DropdownMenuItem>
+  </DropdownMenuContent>
+</DropdownMenu>`,
+  },
+  menubar: {
+    componentName: 'Menubar',
+    importDocs: `
+import {
+  Menubar,
+  MenubarContent,
+  MenubarItem,
+  MenubarMenu,
+  MenubarSeparator,
+  MenubarShortcut,
+  MenubarTrigger,
+} from "/components/ui/menubar"`,
+    usageDocs: `
+<Menubar>
+  <MenubarMenu>
+    <MenubarTrigger>File</MenubarTrigger>
+    <MenubarContent>
+      <MenubarItem>
+        New Tab <MenubarShortcut>⌘T</MenubarShortcut>
+      </MenubarItem>
+      <MenubarItem>New Window</MenubarItem>
+      <MenubarSeparator />
+      <MenubarItem>Share</MenubarItem>
+      <MenubarSeparator />
+      <MenubarItem>Print</MenubarItem>
+    </MenubarContent>
+  </MenubarMenu>
+</Menubar>`,
+  },
+  navigationMenu: {
+    componentName: 'NavigationMenu',
+    importDocs: `
+import {
+  NavigationMenu,
+  NavigationMenuContent,
+  NavigationMenuItem,
+  NavigationMenuLink,
+  NavigationMenuList,
+  NavigationMenuTrigger,
+  navigationMenuTriggerStyle,
+} from "/components/ui/navigation-menu"`,
+    usageDocs: `
+<NavigationMenu>
+  <NavigationMenuList>
+    <NavigationMenuItem>
+      <NavigationMenuTrigger>Item One</NavigationMenuTrigger>
+      <NavigationMenuContent>
+        <NavigationMenuLink>Link</NavigationMenuLink>
+      </NavigationMenuContent>
+    </NavigationMenuItem>
+  </NavigationMenuList>
+</NavigationMenu>`,
+  },
+  popover: {
+    componentName: 'Popover',
+    importDocs: `
+import {
+  Popover,
+  PopoverContent,
+  PopoverTrigger,
+} from "/components/ui/popover"`,
+    usageDocs: `
+<Popover>
+  <PopoverTrigger>Open</PopoverTrigger>
+  <PopoverContent>Place content for the popover here.</PopoverContent>
+</Popover>`,
+  },
+  progress: {
+    componentName: 'Progress',
+    importDocs: 'import { Progress } from "/components/ui/progress"',
+    usageDocs: '<Progress value={33} />',
+  },
+  separator: {
+    componentName: 'Separator',
+    importDocs: 'import { Separator } from "/components/ui/separator"',
+    usageDocs: '<Separator />',
+  },
+  sheet: {
+    componentName: 'Sheet',
+    importDocs: `
+import {
+  Sheet,
+  SheetContent,
+  SheetDescription,
+  SheetHeader,
+  SheetTitle,
+  SheetTrigger,
+} from "/components/ui/sheet"`,
+    usageDocs: `
+<Sheet>
+  <SheetTrigger>Open</SheetTrigger>
+  <SheetContent>
+    <SheetHeader>
+      <SheetTitle>Are you sure absolutely sure?</SheetTitle>
+      <SheetDescription>
+        This action cannot be undone.
+      </SheetDescription>
+    </SheetHeader>
+  </SheetContent>
+</Sheet>`,
+  },
+  skeleton: {
+    componentName: 'Skeleton',
+    importDocs: 'import { Skeleton } from "/components/ui/skeleton"',
+    usageDocs: '<Skeleton className="w-[100px] h-[20px] rounded-full" />',
+  },
+  slider: {
+    componentName: 'Slider',
+    importDocs: 'import { Slider } from "/components/ui/slider"',
+    usageDocs: '<Slider defaultValue={[33]} max={100} step={1} />',
+  },
+  switch: {
+    componentName: 'Switch',
+    importDocs: 'import { Switch } from "/components/ui/switch"',
+    usageDocs: '<Switch />',
+  },
+  table: {
+    componentName: 'Table',
+    importDocs: `
+import {
+  Table,
+  TableBody,
+  TableCaption,
+  TableCell,
+  TableHead,
+  TableHeader,
+  TableRow,
+} from "/components/ui/table"`,
+    usageDocs: `
+<Table>
+  <TableCaption>A list of your recent invoices.</TableCaption>
+  <TableHeader>
+    <TableRow>
+      <TableHead className="w-[100px]">Invoice</TableHead>
+      <TableHead>Status</TableHead>
+      <TableHead>Method</TableHead>
+      <TableHead className="text-right">Amount</TableHead>
+    </TableRow>
+  </TableHeader>
+  <TableBody>
+    <TableRow>
+      <TableCell className="font-medium">INV001</TableCell>
+      <TableCell>Paid</TableCell>
+      <TableCell>Credit Card</TableCell>
+      <TableCell className="text-right">$250.00</TableCell>
+    </TableRow>
+  </TableBody>
+</Table>`,
+  },
+  tabs: {
+    componentName: 'Tabs',
+    importDocs: `
+import {
+  Tabs,
+  TabsContent,
+  TabsList,
+  TabsTrigger,
+} from "/components/ui/tabs"`,
+    usageDocs: `
+<Tabs defaultValue="account" className="w-[400px]">
+  <TabsList>
+    <TabsTrigger value="account">Account</TabsTrigger>
+    <TabsTrigger value="password">Password</TabsTrigger>
+  </TabsList>
+  <TabsContent value="account">Make changes to your account here.</TabsContent>
+  <TabsContent value="password">Change your password here.</TabsContent>
+</Tabs>`,
+  },
+  toast: {
+    componentName: 'Toast',
+    importDocs: `
+import { useToast } from "/components/ui/use-toast"
+import { Button } from "/components/ui/button"`,
+    usageDocs: `
+export function ToastDemo() {
+  const { toast } = useToast()
+  return (
+    <Button
+      onClick={() => {
+        toast({
+          title: "Scheduled: Catch up",
+          description: "Friday, February 10, 2023 at 5:57 PM",
+        })
+      }}
+    >
+      Show Toast
+    </Button>
+  )
+}`,
+  },
+  toggle: {
+    componentName: 'Toggle',
+    importDocs: 'import { Toggle } from "/components/ui/toggle"',
+    usageDocs: '<Toggle>Toggle</Toggle>',
+  },
+  tooltip: {
+    componentName: 'Tooltip',
+    importDocs: `
+import {
+  Tooltip,
+  TooltipContent,
+  TooltipProvider,
+  TooltipTrigger,
+} from "/components/ui/tooltip"`,
+    usageDocs: `
+<TooltipProvider>
+  <Tooltip>
+    <TooltipTrigger>Hover</TooltipTrigger>
+    <TooltipContent>
+      <p>Add to library</p>
+    </TooltipContent>
+  </Tooltip>
+</TooltipProvider>`,
+  },
+};
+
+const components = Object.assign({}, essentialComponents, extraComponents);
+
+module.exports = {
+  components,
+};

api/app/clients/prompts/shadcn-docs/generate.js

@@ -0,0 +1,50 @@
+const dedent = require('dedent');
+
+/**
+ * Generate system prompt for AI-assisted React component creation
+ * @param {Object} options - Configuration options
+ * @param {Object} options.components - Documentation for shadcn components
+ * @param {boolean} [options.useXML=false] - Whether to use XML-style formatting for component instructions
+ * @returns {string} The generated system prompt
+ */
+function generateShadcnPrompt(options) {
+  const { components, useXML = false } = options;
+
+  let systemPrompt = dedent`
+    ## Additional Artifact Instructions for React Components: "application/vnd.react"
+    
+    There are some prestyled components (primitives) available for use. Please use your best judgement to use any of these components if the app calls for one.
+
+    Here are the components that are available, along with how to import them, and how to use them:
+
+    ${Object.values(components)
+    .map((component) => {
+      if (useXML) {
+        return dedent`
+            <component>
+              <name>${component.componentName}</name>
+              <import-instructions>${component.importDocs}</import-instructions>
+              <usage-instructions>${component.usageDocs}</usage-instructions>
+            </component>
+          `;
+      } else {
+        return dedent`
+            # ${component.componentName}
+
+            ## Import Instructions
+            ${component.importDocs}
+
+            ## Usage Instructions
+            ${component.usageDocs}
+          `;
+      }
+    })
+    .join('\n\n')}
+  `;
+
+  return systemPrompt;
+}
+
+module.exports = {
+  generateShadcnPrompt,
+};

api/app/clients/prompts/summaryPrompts.js

@@ -1,4 +1,4 @@
-const { PromptTemplate } = require('langchain/prompts');
+const { PromptTemplate } = require('@langchain/core/prompts');
 /*
  * Without `{summary}` and `{new_lines}`, token count is 98
  * We are counting this towards the max context tokens for summaries, +3 for the assistant label (101)

api/app/clients/prompts/titlePrompts.js

@@ -2,7 +2,7 @@ const {
   ChatPromptTemplate,
   SystemMessagePromptTemplate,
   HumanMessagePromptTemplate,
-} = require('langchain/prompts');
+} = require('@langchain/core/prompts');
 
 const langPrompt = new ChatPromptTemplate({
   promptMessages: [
@@ -27,7 +27,110 @@ ${convo}`,
   return titlePrompt;
 };
 
+const titleInstruction =
+  'a concise, 5-word-or-less title for the conversation, using its same language, with no punctuation. Apply title case conventions appropriate for the language. Never directly mention the language name or the word "title"';
+const titleFunctionPrompt = `In this environment you have access to a set of tools you can use to generate the conversation title.
+  
+You may call them like this:
+<function_calls>
+<invoke>
+<tool_name>$TOOL_NAME</tool_name>
+<parameters>
+<$PARAMETER_NAME>$PARAMETER_VALUE</$PARAMETER_NAME>
+...
+</parameters>
+</invoke>
+</function_calls>
+
+Here are the tools available:
+<tools>
+<tool_description>
+<tool_name>submit_title</tool_name>
+<description>
+Submit a brief title in the conversation's language, following the parameter description closely.
+</description>
+<parameters>
+<parameter>
+<name>title</name>
+<type>string</type>
+<description>${titleInstruction}</description>
+</parameter>
+</parameters>
+</tool_description>
+</tools>`;
+
+const genTranslationPrompt = (
+  translationPrompt,
+) => `In this environment you have access to a set of tools you can use to translate text.
+  
+You may call them like this:
+<function_calls>
+<invoke>
+<tool_name>$TOOL_NAME</tool_name>
+<parameters>
+<$PARAMETER_NAME>$PARAMETER_VALUE</$PARAMETER_NAME>
+...
+</parameters>
+</invoke>
+</function_calls>
+
+Here are the tools available:
+<tools>
+<tool_description>
+<tool_name>submit_translation</tool_name>
+<description>
+Submit a translation in the target language, following the parameter description and its language closely.
+</description>
+<parameters>
+<parameter>
+<name>translation</name>
+<type>string</type>
+<description>${translationPrompt}
+ONLY include the generated translation without quotations, nor its related key</description>
+</parameter>
+</parameters>
+</tool_description>
+</tools>`;
+
+/**
+ * Parses specified parameter from the provided prompt.
+ * @param {string} prompt - The prompt containing the desired parameter.
+ * @param {string} paramName - The name of the parameter to extract.
+ * @returns {string} The parsed parameter's value or a default value if not found.
+ */
+function parseParamFromPrompt(prompt, paramName) {
+  // Handle null/undefined prompt
+  if (!prompt) {
+    return `No ${paramName} provided`;
+  }
+
+  // Try original format first: <title>value</title>
+  const simpleRegex = new RegExp(`<${paramName}>(.*?)</${paramName}>`, 's');
+  const simpleMatch = prompt.match(simpleRegex);
+
+  if (simpleMatch) {
+    return simpleMatch[1].trim();
+  }
+
+  // Try parameter format: <parameter name="title">value</parameter>
+  const paramRegex = new RegExp(`<parameter name="${paramName}">(.*?)</parameter>`, 's');
+  const paramMatch = prompt.match(paramRegex);
+
+  if (paramMatch) {
+    return paramMatch[1].trim();
+  }
+
+  if (prompt && prompt.length) {
+    return `NO TOOL INVOCATION: ${prompt}`;
+  }
+  return `No ${paramName} provided`;
+}
+
 module.exports = {
   langPrompt,
+  titleInstruction,
   createTitlePrompt,
+  titleFunctionPrompt,
+  parseParamFromPrompt,
+  genTranslationPrompt,
 };

api/app/clients/prompts/titlePrompts.spec.js

@@ -0,0 +1,73 @@
+const { parseParamFromPrompt } = require('./titlePrompts');
+describe('parseParamFromPrompt', () => {
+  // Original simple format tests
+  test('extracts parameter from simple format', () => {
+    const prompt = '<title>Simple Title</title>';
+    expect(parseParamFromPrompt(prompt, 'title')).toBe('Simple Title');
+  });
+
+  // Parameter format tests
+  test('extracts parameter from parameter format', () => {
+    const prompt =
+      '<function_calls> <invoke name="submit_title"> <parameter name="title">Complex Title</parameter> </invoke>';
+    expect(parseParamFromPrompt(prompt, 'title')).toBe('Complex Title');
+  });
+
+  // Edge cases and error handling
+  test('returns NO TOOL INVOCATION message for non-matching content', () => {
+    const prompt = 'Some random text without parameters';
+    expect(parseParamFromPrompt(prompt, 'title')).toBe(
+      'NO TOOL INVOCATION: Some random text without parameters',
+    );
+  });
+
+  test('returns default message for empty prompt', () => {
+    expect(parseParamFromPrompt('', 'title')).toBe('No title provided');
+  });
+
+  test('returns default message for null prompt', () => {
+    expect(parseParamFromPrompt(null, 'title')).toBe('No title provided');
+  });
+
+  // Multiple parameter tests
+  test('works with different parameter names', () => {
+    const prompt = '<name>John Doe</name>';
+    expect(parseParamFromPrompt(prompt, 'name')).toBe('John Doe');
+  });
+
+  test('handles multiline content', () => {
+    const prompt = `<parameter name="description">This is a
+    multiline
+    description</parameter>`;
+    expect(parseParamFromPrompt(prompt, 'description')).toBe(
+      'This is a\n    multiline\n    description',
+    );
+  });
+
+  // Whitespace handling
+  test('trims whitespace from extracted content', () => {
+    const prompt = '<title>  Padded Title  </title>';
+    expect(parseParamFromPrompt(prompt, 'title')).toBe('Padded Title');
+  });
+
+  test('handles whitespace in parameter format', () => {
+    const prompt = '<parameter name="title">  Padded Parameter Title  </parameter>';
+    expect(parseParamFromPrompt(prompt, 'title')).toBe('Padded Parameter Title');
+  });
+
+  // Invalid format tests
+  test('handles malformed tags', () => {
+    const prompt = '<title>Incomplete Tag';
+    expect(parseParamFromPrompt(prompt, 'title')).toBe('NO TOOL INVOCATION: <title>Incomplete Tag');
+  });
+
+  test('handles empty tags', () => {
+    const prompt = '<title></title>';
+    expect(parseParamFromPrompt(prompt, 'title')).toBe('');
+  });
+
+  test('handles empty parameter tags', () => {
+    const prompt = '<parameter name="title"></parameter>';
+    expect(parseParamFromPrompt(prompt, 'title')).toBe('');
+  });
+});

api/app/clients/prompts/truncate.js

@@ -0,0 +1,115 @@
+const MAX_CHAR = 255;
+
+/**
+ * Truncates a given text to a specified maximum length, appending ellipsis and a notification
+ * if the original text exceeds the maximum length.
+ *
+ * @param {string} text - The text to be truncated.
+ * @param {number} [maxLength=MAX_CHAR] - The maximum length of the text after truncation. Defaults to MAX_CHAR.
+ * @returns {string} The truncated text if the original text length exceeds maxLength, otherwise returns the original text.
+ */
+function truncateText(text, maxLength = MAX_CHAR) {
+  if (text.length > maxLength) {
+    return `${text.slice(0, maxLength)}... [text truncated for brevity]`;
+  }
+  return text;
+}
+
+/**
+ * Truncates a given text to a specified maximum length by showing the first half and the last half of the text,
+ * separated by ellipsis. This method ensures the output does not exceed the maximum length, including the addition
+ * of ellipsis and notification if the original text exceeds the maximum length.
+ *
+ * @param {string} text - The text to be truncated.
+ * @param {number} [maxLength=MAX_CHAR] - The maximum length of the output text after truncation. Defaults to MAX_CHAR.
+ * @returns {string} The truncated text showing the first half and the last half, or the original text if it does not exceed maxLength.
+ */
+function smartTruncateText(text, maxLength = MAX_CHAR) {
+  const ellipsis = '...';
+  const notification = ' [text truncated for brevity]';
+  const halfMaxLength = Math.floor((maxLength - ellipsis.length - notification.length) / 2);
+
+  if (text.length > maxLength) {
+    const startLastHalf = text.length - halfMaxLength;
+    return `${text.slice(0, halfMaxLength)}${ellipsis}${text.slice(startLastHalf)}${notification}`;
+  }
+
+  return text;
+}
+
+/**
+ * @param {TMessage[]} _messages
+ * @param {number} maxContextTokens
+ * @param {function({role: string, content: TMessageContent[]}): number} getTokenCountForMessage
+ *
+ * @returns {{
+ *  dbMessages: TMessage[],
+ * editedIndices: number[]
+ * }}
+ */
+function truncateToolCallOutputs(_messages, maxContextTokens, getTokenCountForMessage) {
+  const THRESHOLD_PERCENTAGE = 0.5;
+  const targetTokenLimit = maxContextTokens * THRESHOLD_PERCENTAGE;
+
+  let currentTokenCount = 3;
+  const messages = [..._messages];
+  const processedMessages = [];
+  let currentIndex = messages.length;
+  const editedIndices = new Set();
+  while (messages.length > 0) {
+    currentIndex--;
+    const message = messages.pop();
+    currentTokenCount += message.tokenCount;
+    if (currentTokenCount < targetTokenLimit) {
+      processedMessages.push(message);
+      continue;
+    }
+
+    if (!message.content || !Array.isArray(message.content)) {
+      processedMessages.push(message);
+      continue;
+    }
+
+    const toolCallIndices = message.content
+      .map((item, index) => (item.type === 'tool_call' ? index : -1))
+      .filter((index) => index !== -1)
+      .reverse();
+
+    if (toolCallIndices.length === 0) {
+      processedMessages.push(message);
+      continue;
+    }
+
+    const newContent = [...message.content];
+
+    // Truncate all tool outputs since we're over threshold
+    for (const index of toolCallIndices) {
+      const toolCall = newContent[index].tool_call;
+      if (!toolCall || !toolCall.output) {
+        continue;
+      }
+
+      editedIndices.add(currentIndex);
+
+      newContent[index] = {
+        ...newContent[index],
+        tool_call: {
+          ...toolCall,
+          output: '[OUTPUT_OMITTED_FOR_BREVITY]',
+        },
+      };
+    }
+
+    const truncatedMessage = {
+      ...message,
+      content: newContent,
+      tokenCount: getTokenCountForMessage({ role: 'assistant', content: newContent }),
+    };
+
+    processedMessages.push(truncatedMessage);
+  }
+
+  return { dbMessages: processedMessages.reverse(), editedIndices: Array.from(editedIndices) };
+}
+
+module.exports = { truncateText, smartTruncateText, truncateToolCallOutputs };

api/app/clients/prompts/truncateText.js

@@ -1,10 +0,0 @@
-const MAX_CHAR = 255;
-
-function truncateText(text) {
-  if (text.length > MAX_CHAR) {
-    return `${text.slice(0, MAX_CHAR)}... [text truncated for brevity]`;
-  }
-  return text;
-}
-
-module.exports = truncateText;

api/app/clients/specs/AnthropicClient.test.js

@@ -1,4 +1,7 @@
-const AnthropicClient = require('../AnthropicClient');
+const { SplitStreamHandler } = require('@librechat/agents');
+const { anthropicSettings } = require('librechat-data-provider');
+const AnthropicClient = require('~/app/clients/AnthropicClient');
+
 const HUMAN_PROMPT = '\n\nHuman:';
 const AI_PROMPT = '\n\nAssistant:';
 
@@ -22,7 +25,7 @@ describe('AnthropicClient', () => {
     const options = {
       modelOptions: {
         model,
-        temperature: 0.7,
+        temperature: anthropicSettings.temperature.default,
       },
     };
     client = new AnthropicClient('test-api-key');
@@ -33,7 +36,42 @@ describe('AnthropicClient', () => {
     it('should set the options correctly', () => {
       expect(client.apiKey).toBe('test-api-key');
       expect(client.modelOptions.model).toBe(model);
-      expect(client.modelOptions.temperature).toBe(0.7);
+      expect(client.modelOptions.temperature).toBe(anthropicSettings.temperature.default);
+    });
+
+    it('should set legacy maxOutputTokens for non-Claude-3 models', () => {
+      const client = new AnthropicClient('test-api-key');
+      client.setOptions({
+        modelOptions: {
+          model: 'claude-2',
+          maxOutputTokens: anthropicSettings.maxOutputTokens.default,
+        },
+      });
+      expect(client.modelOptions.maxOutputTokens).toBe(
+        anthropicSettings.legacy.maxOutputTokens.default,
+      );
+    });
+    it('should not set maxOutputTokens if not provided', () => {
+      const client = new AnthropicClient('test-api-key');
+      client.setOptions({
+        modelOptions: {
+          model: 'claude-3',
+        },
+      });
+      expect(client.modelOptions.maxOutputTokens).toBeUndefined();
+    });
+
+    it('should not set legacy maxOutputTokens for Claude-3 models', () => {
+      const client = new AnthropicClient('test-api-key');
+      client.setOptions({
+        modelOptions: {
+          model: 'claude-3-opus-20240229',
+          maxOutputTokens: anthropicSettings.legacy.maxOutputTokens.default,
+        },
+      });
+      expect(client.modelOptions.maxOutputTokens).toBe(
+        anthropicSettings.legacy.maxOutputTokens.default,
+      );
     });
   });
 
@@ -136,4 +174,510 @@ describe('AnthropicClient', () => {
       expect(prompt).toContain('You are Claude-2');
     });
   });
+
+  describe('getClient', () => {
+    it('should set legacy maxOutputTokens for non-Claude-3 models', () => {
+      const client = new AnthropicClient('test-api-key');
+      client.setOptions({
+        modelOptions: {
+          model: 'claude-2',
+          maxOutputTokens: anthropicSettings.legacy.maxOutputTokens.default,
+        },
+      });
+      expect(client.modelOptions.maxOutputTokens).toBe(
+        anthropicSettings.legacy.maxOutputTokens.default,
+      );
+    });
+
+    it('should not set legacy maxOutputTokens for Claude-3 models', () => {
+      const client = new AnthropicClient('test-api-key');
+      client.setOptions({
+        modelOptions: {
+          model: 'claude-3-opus-20240229',
+          maxOutputTokens: anthropicSettings.legacy.maxOutputTokens.default,
+        },
+      });
+      expect(client.modelOptions.maxOutputTokens).toBe(
+        anthropicSettings.legacy.maxOutputTokens.default,
+      );
+    });
+
+    it('should add "max-tokens" & "prompt-caching" beta header for claude-3-5-sonnet model', () => {
+      const client = new AnthropicClient('test-api-key');
+      const modelOptions = {
+        model: 'claude-3-5-sonnet-20241022',
+      };
+      client.setOptions({ modelOptions, promptCache: true });
+      const anthropicClient = client.getClient(modelOptions);
+      expect(anthropicClient._options.defaultHeaders).toBeDefined();
+      expect(anthropicClient._options.defaultHeaders).toHaveProperty('anthropic-beta');
+      expect(anthropicClient._options.defaultHeaders['anthropic-beta']).toBe(
+        'max-tokens-3-5-sonnet-2024-07-15,prompt-caching-2024-07-31',
+      );
+    });
+
+    it('should add "prompt-caching" beta header for claude-3-haiku model', () => {
+      const client = new AnthropicClient('test-api-key');
+      const modelOptions = {
+        model: 'claude-3-haiku-2028',
+      };
+      client.setOptions({ modelOptions, promptCache: true });
+      const anthropicClient = client.getClient(modelOptions);
+      expect(anthropicClient._options.defaultHeaders).toBeDefined();
+      expect(anthropicClient._options.defaultHeaders).toHaveProperty('anthropic-beta');
+      expect(anthropicClient._options.defaultHeaders['anthropic-beta']).toBe(
+        'prompt-caching-2024-07-31',
+      );
+    });
+
+    it('should add "prompt-caching" beta header for claude-3-opus model', () => {
+      const client = new AnthropicClient('test-api-key');
+      const modelOptions = {
+        model: 'claude-3-opus-2028',
+      };
+      client.setOptions({ modelOptions, promptCache: true });
+      const anthropicClient = client.getClient(modelOptions);
+      expect(anthropicClient._options.defaultHeaders).toBeDefined();
+      expect(anthropicClient._options.defaultHeaders).toHaveProperty('anthropic-beta');
+      expect(anthropicClient._options.defaultHeaders['anthropic-beta']).toBe(
+        'prompt-caching-2024-07-31',
+      );
+    });
+
+    it('should not add beta header for claude-3-5-sonnet-latest model', () => {
+      const client = new AnthropicClient('test-api-key');
+      const modelOptions = {
+        model: 'anthropic/claude-3-5-sonnet-latest',
+      };
+      client.setOptions({ modelOptions, promptCache: true });
+      const anthropicClient = client.getClient(modelOptions);
+      expect(anthropicClient.defaultHeaders).not.toHaveProperty('anthropic-beta');
+    });
+
+    it('should not add beta header for other models', () => {
+      const client = new AnthropicClient('test-api-key');
+      client.setOptions({
+        modelOptions: {
+          model: 'claude-2',
+        },
+      });
+      const anthropicClient = client.getClient();
+      expect(anthropicClient.defaultHeaders).not.toHaveProperty('anthropic-beta');
+    });
+  });
+
+  describe('calculateCurrentTokenCount', () => {
+    let client;
+
+    beforeEach(() => {
+      client = new AnthropicClient('test-api-key');
+    });
+
+    it('should calculate correct token count when usage is provided', () => {
+      const tokenCountMap = {
+        msg1: 10,
+        msg2: 20,
+        currentMsg: 30,
+      };
+      const currentMessageId = 'currentMsg';
+      const usage = {
+        input_tokens: 70,
+        output_tokens: 50,
+      };
+
+      const result = client.calculateCurrentTokenCount({ tokenCountMap, currentMessageId, usage });
+
+      expect(result).toBe(40); // 70 - (10 + 20) = 40
+    });
+
+    it('should return original estimate if calculation results in negative value', () => {
+      const tokenCountMap = {
+        msg1: 40,
+        msg2: 50,
+        currentMsg: 30,
+      };
+      const currentMessageId = 'currentMsg';
+      const usage = {
+        input_tokens: 80,
+        output_tokens: 50,
+      };
+
+      const result = client.calculateCurrentTokenCount({ tokenCountMap, currentMessageId, usage });
+
+      expect(result).toBe(30); // Original estimate
+    });
+
+    it('should handle cache creation and read input tokens', () => {
+      const tokenCountMap = {
+        msg1: 10,
+        msg2: 20,
+        currentMsg: 30,
+      };
+      const currentMessageId = 'currentMsg';
+      const usage = {
+        input_tokens: 50,
+        cache_creation_input_tokens: 10,
+        cache_read_input_tokens: 20,
+        output_tokens: 40,
+      };
+
+      const result = client.calculateCurrentTokenCount({ tokenCountMap, currentMessageId, usage });
+
+      expect(result).toBe(50); // (50 + 10 + 20) - (10 + 20) = 50
+    });
+
+    it('should handle missing usage properties', () => {
+      const tokenCountMap = {
+        msg1: 10,
+        msg2: 20,
+        currentMsg: 30,
+      };
+      const currentMessageId = 'currentMsg';
+      const usage = {
+        output_tokens: 40,
+      };
+
+      const result = client.calculateCurrentTokenCount({ tokenCountMap, currentMessageId, usage });
+
+      expect(result).toBe(30); // Original estimate
+    });
+
+    it('should handle empty tokenCountMap', () => {
+      const tokenCountMap = {};
+      const currentMessageId = 'currentMsg';
+      const usage = {
+        input_tokens: 50,
+        output_tokens: 40,
+      };
+
+      const result = client.calculateCurrentTokenCount({ tokenCountMap, currentMessageId, usage });
+
+      expect(result).toBe(50);
+      expect(Number.isNaN(result)).toBe(false);
+    });
+
+    it('should handle zero values in usage', () => {
+      const tokenCountMap = {
+        msg1: 10,
+        currentMsg: 20,
+      };
+      const currentMessageId = 'currentMsg';
+      const usage = {
+        input_tokens: 0,
+        cache_creation_input_tokens: 0,
+        cache_read_input_tokens: 0,
+        output_tokens: 0,
+      };
+
+      const result = client.calculateCurrentTokenCount({ tokenCountMap, currentMessageId, usage });
+
+      expect(result).toBe(20); // Should return original estimate
+      expect(Number.isNaN(result)).toBe(false);
+    });
+
+    it('should handle undefined usage', () => {
+      const tokenCountMap = {
+        msg1: 10,
+        currentMsg: 20,
+      };
+      const currentMessageId = 'currentMsg';
+      const usage = undefined;
+
+      const result = client.calculateCurrentTokenCount({ tokenCountMap, currentMessageId, usage });
+
+      expect(result).toBe(20); // Should return original estimate
+      expect(Number.isNaN(result)).toBe(false);
+    });
+
+    it('should handle non-numeric values in tokenCountMap', () => {
+      const tokenCountMap = {
+        msg1: 'ten',
+        currentMsg: 20,
+      };
+      const currentMessageId = 'currentMsg';
+      const usage = {
+        input_tokens: 30,
+        output_tokens: 10,
+      };
+
+      const result = client.calculateCurrentTokenCount({ tokenCountMap, currentMessageId, usage });
+
+      expect(result).toBe(30); // Should return 30 (input_tokens) - 0 (ignored 'ten') = 30
+      expect(Number.isNaN(result)).toBe(false);
+    });
+  });
+
+  describe('maxOutputTokens handling for different models', () => {
+    it('should not cap maxOutputTokens for Claude 3.5 Sonnet models', () => {
+      const client = new AnthropicClient('test-api-key');
+      const highTokenValue = anthropicSettings.legacy.maxOutputTokens.default * 10;
+
+      client.setOptions({
+        modelOptions: {
+          model: 'claude-3-5-sonnet',
+          maxOutputTokens: highTokenValue,
+        },
+      });
+
+      expect(client.modelOptions.maxOutputTokens).toBe(highTokenValue);
+
+      // Test with decimal notation
+      client.setOptions({
+        modelOptions: {
+          model: 'claude-3.5-sonnet',
+          maxOutputTokens: highTokenValue,
+        },
+      });
+
+      expect(client.modelOptions.maxOutputTokens).toBe(highTokenValue);
+    });
+
+    it('should not cap maxOutputTokens for Claude 3.7 models', () => {
+      const client = new AnthropicClient('test-api-key');
+      const highTokenValue = anthropicSettings.legacy.maxOutputTokens.default * 2;
+
+      client.setOptions({
+        modelOptions: {
+          model: 'claude-3-7-sonnet',
+          maxOutputTokens: highTokenValue,
+        },
+      });
+
+      expect(client.modelOptions.maxOutputTokens).toBe(highTokenValue);
+
+      // Test with decimal notation
+      client.setOptions({
+        modelOptions: {
+          model: 'claude-3.7-sonnet',
+          maxOutputTokens: highTokenValue,
+        },
+      });
+
+      expect(client.modelOptions.maxOutputTokens).toBe(highTokenValue);
+    });
+
+    it('should cap maxOutputTokens for Claude 3.5 Haiku models', () => {
+      const client = new AnthropicClient('test-api-key');
+      const highTokenValue = anthropicSettings.legacy.maxOutputTokens.default * 2;
+
+      client.setOptions({
+        modelOptions: {
+          model: 'claude-3-5-haiku',
+          maxOutputTokens: highTokenValue,
+        },
+      });
+
+      expect(client.modelOptions.maxOutputTokens).toBe(
+        anthropicSettings.legacy.maxOutputTokens.default,
+      );
+
+      // Test with decimal notation
+      client.setOptions({
+        modelOptions: {
+          model: 'claude-3.5-haiku',
+          maxOutputTokens: highTokenValue,
+        },
+      });
+
+      expect(client.modelOptions.maxOutputTokens).toBe(
+        anthropicSettings.legacy.maxOutputTokens.default,
+      );
+    });
+
+    it('should cap maxOutputTokens for Claude 3 Haiku and Opus models', () => {
+      const client = new AnthropicClient('test-api-key');
+      const highTokenValue = anthropicSettings.legacy.maxOutputTokens.default * 2;
+
+      // Test haiku
+      client.setOptions({
+        modelOptions: {
+          model: 'claude-3-haiku',
+          maxOutputTokens: highTokenValue,
+        },
+      });
+
+      expect(client.modelOptions.maxOutputTokens).toBe(
+        anthropicSettings.legacy.maxOutputTokens.default,
+      );
+
+      // Test opus
+      client.setOptions({
+        modelOptions: {
+          model: 'claude-3-opus',
+          maxOutputTokens: highTokenValue,
+        },
+      });
+
+      expect(client.modelOptions.maxOutputTokens).toBe(
+        anthropicSettings.legacy.maxOutputTokens.default,
+      );
+    });
+  });
+
+  describe('topK/topP parameters for different models', () => {
+    beforeEach(() => {
+      // Mock the SplitStreamHandler
+      jest.spyOn(SplitStreamHandler.prototype, 'handle').mockImplementation(() => {});
+    });
+
+    afterEach(() => {
+      jest.restoreAllMocks();
+    });
+
+    it('should include top_k and top_p parameters for non-claude-3.7 models', async () => {
+      const client = new AnthropicClient('test-api-key');
+
+      // Create a mock async generator function
+      async function* mockAsyncGenerator() {
+        yield { type: 'message_start', message: { usage: {} } };
+        yield { delta: { text: 'Test response' } };
+        yield { type: 'message_delta', usage: {} };
+      }
+
+      // Mock createResponse to return the async generator
+      jest.spyOn(client, 'createResponse').mockImplementation(() => {
+        return mockAsyncGenerator();
+      });
+
+      client.setOptions({
+        modelOptions: {
+          model: 'claude-3-opus',
+          temperature: 0.7,
+          topK: 10,
+          topP: 0.9,
+        },
+      });
+
+      // Mock getClient to capture the request options
+      let capturedOptions = null;
+      jest.spyOn(client, 'getClient').mockImplementation((options) => {
+        capturedOptions = options;
+        return {};
+      });
+
+      const payload = [{ role: 'user', content: 'Test message' }];
+      await client.sendCompletion(payload, {});
+
+      // Check the options passed to getClient
+      expect(capturedOptions).toHaveProperty('top_k', 10);
+      expect(capturedOptions).toHaveProperty('top_p', 0.9);
+    });
+
+    it('should include top_k and top_p parameters for claude-3-5-sonnet models', async () => {
+      const client = new AnthropicClient('test-api-key');
+
+      // Create a mock async generator function
+      async function* mockAsyncGenerator() {
+        yield { type: 'message_start', message: { usage: {} } };
+        yield { delta: { text: 'Test response' } };
+        yield { type: 'message_delta', usage: {} };
+      }
+
+      // Mock createResponse to return the async generator
+      jest.spyOn(client, 'createResponse').mockImplementation(() => {
+        return mockAsyncGenerator();
+      });
+
+      client.setOptions({
+        modelOptions: {
+          model: 'claude-3-5-sonnet',
+          temperature: 0.7,
+          topK: 10,
+          topP: 0.9,
+        },
+      });
+
+      // Mock getClient to capture the request options
+      let capturedOptions = null;
+      jest.spyOn(client, 'getClient').mockImplementation((options) => {
+        capturedOptions = options;
+        return {};
+      });
+
+      const payload = [{ role: 'user', content: 'Test message' }];
+      await client.sendCompletion(payload, {});
+
+      // Check the options passed to getClient
+      expect(capturedOptions).toHaveProperty('top_k', 10);
+      expect(capturedOptions).toHaveProperty('top_p', 0.9);
+    });
+
+    it('should not include top_k and top_p parameters for claude-3-7-sonnet models', async () => {
+      const client = new AnthropicClient('test-api-key');
+
+      // Create a mock async generator function
+      async function* mockAsyncGenerator() {
+        yield { type: 'message_start', message: { usage: {} } };
+        yield { delta: { text: 'Test response' } };
+        yield { type: 'message_delta', usage: {} };
+      }
+
+      // Mock createResponse to return the async generator
+      jest.spyOn(client, 'createResponse').mockImplementation(() => {
+        return mockAsyncGenerator();
+      });
+
+      client.setOptions({
+        modelOptions: {
+          model: 'claude-3-7-sonnet',
+          temperature: 0.7,
+          topK: 10,
+          topP: 0.9,
+        },
+      });
+
+      // Mock getClient to capture the request options
+      let capturedOptions = null;
+      jest.spyOn(client, 'getClient').mockImplementation((options) => {
+        capturedOptions = options;
+        return {};
+      });
+
+      const payload = [{ role: 'user', content: 'Test message' }];
+      await client.sendCompletion(payload, {});
+
+      // Check the options passed to getClient
+      expect(capturedOptions).not.toHaveProperty('top_k');
+      expect(capturedOptions).not.toHaveProperty('top_p');
+    });
+
+    it('should not include top_k and top_p parameters for models with decimal notation (claude-3.7)', async () => {
+      const client = new AnthropicClient('test-api-key');
+
+      // Create a mock async generator function
+      async function* mockAsyncGenerator() {
+        yield { type: 'message_start', message: { usage: {} } };
+        yield { delta: { text: 'Test response' } };
+        yield { type: 'message_delta', usage: {} };
+      }
+
+      // Mock createResponse to return the async generator
+      jest.spyOn(client, 'createResponse').mockImplementation(() => {
+        return mockAsyncGenerator();
+      });
+
+      client.setOptions({
+        modelOptions: {
+          model: 'claude-3.7-sonnet',
+          temperature: 0.7,
+          topK: 10,
+          topP: 0.9,
+        },
+      });
+
+      // Mock getClient to capture the request options
+      let capturedOptions = null;
+      jest.spyOn(client, 'getClient').mockImplementation((options) => {
+        capturedOptions = options;
+        return {};
+      });
+
+      const payload = [{ role: 'user', content: 'Test message' }];
+      await client.sendCompletion(payload, {});
+
+      // Check the options passed to getClient
+      expect(capturedOptions).not.toHaveProperty('top_k');
+      expect(capturedOptions).not.toHaveProperty('top_p');
+    });
+  });
 });

api/app/clients/specs/BaseClient.test.js

@@ -1,21 +1,36 @@
+const { Constants } = require('librechat-data-provider');
 const { initializeFakeClient } = require('./FakeClient');
 
-jest.mock('../../../lib/db/connectDb');
-jest.mock('../../../models', () => {
-  return function () {
-    return {
-      save: jest.fn(),
-      deleteConvos: jest.fn(),
-      getConvo: jest.fn(),
-      getMessages: jest.fn(),
-      saveMessage: jest.fn(),
-      updateMessage: jest.fn(),
-      saveConvo: jest.fn(),
-    };
-  };
-});
+jest.mock('~/lib/db/connectDb');
+jest.mock('~/models', () => ({
+  User: jest.fn(),
+  Key: jest.fn(),
+  Session: jest.fn(),
+  Balance: jest.fn(),
+  Transaction: jest.fn(),
+  getMessages: jest.fn().mockResolvedValue([]),
+  saveMessage: jest.fn(),
+  updateMessage: jest.fn(),
+  deleteMessagesSince: jest.fn(),
+  deleteMessages: jest.fn(),
+  getConvoTitle: jest.fn(),
+  getConvo: jest.fn(),
+  saveConvo: jest.fn(),
+  deleteConvos: jest.fn(),
+  getPreset: jest.fn(),
+  getPresets: jest.fn(),
+  savePreset: jest.fn(),
+  deletePresets: jest.fn(),
+  findFileById: jest.fn(),
+  createFile: jest.fn(),
+  updateFile: jest.fn(),
+  deleteFile: jest.fn(),
+  deleteFiles: jest.fn(),
+  getFiles: jest.fn(),
+  updateFileUsage: jest.fn(),
+}));
 
-jest.mock('langchain/chat_models/openai', () => {
+jest.mock('@langchain/openai', () => {
   return {
     ChatOpenAI: jest.fn().mockImplementation(() => {
       return {};
@@ -46,7 +61,7 @@ describe('BaseClient', () => {
   const options = {
     // debug: true,
     modelOptions: {
-      model: 'gpt-3.5-turbo',
+      model: 'gpt-4o-mini',
       temperature: 0,
     },
   };
@@ -73,6 +88,19 @@ describe('BaseClient', () => {
     const messages = [{ content: 'Hello' }, { content: 'How are you?' }, { content: 'Goodbye' }];
     const instructions = { content: 'Please respond to the question.' };
     const result = TestClient.addInstructions(messages, instructions);
+    const expected = [
+      { content: 'Please respond to the question.' },
+      { content: 'Hello' },
+      { content: 'How are you?' },
+      { content: 'Goodbye' },
+    ];
+    expect(result).toEqual(expected);
+  });
+
+  test('returns the input messages with instructions properly added when addInstructions() with legacy flag', () => {
+    const messages = [{ content: 'Hello' }, { content: 'How are you?' }, { content: 'Goodbye' }];
+    const instructions = { content: 'Please respond to the question.' };
+    const result = TestClient.addInstructions(messages, instructions, true);
     const expected = [
       { content: 'Hello' },
       { content: 'How are you?' },
@@ -131,7 +159,7 @@ describe('BaseClient', () => {
       expectedMessagesToRefine?.[expectedMessagesToRefine.length - 1] ?? {};
     const expectedIndex = messages.findIndex((msg) => msg.content === lastExpectedMessage?.content);
 
-    const result = await TestClient.getMessagesWithinTokenLimit(messages);
+    const result = await TestClient.getMessagesWithinTokenLimit({ messages });
 
     expect(result.context).toEqual(expectedContext);
     expect(result.summaryIndex).toEqual(expectedIndex);
@@ -167,7 +195,7 @@ describe('BaseClient', () => {
       expectedMessagesToRefine?.[expectedMessagesToRefine.length - 1] ?? {};
     const expectedIndex = messages.findIndex((msg) => msg.content === lastExpectedMessage?.content);
 
-    const result = await TestClient.getMessagesWithinTokenLimit(messages);
+    const result = await TestClient.getMessagesWithinTokenLimit({ messages });
 
     expect(result.context).toEqual(expectedContext);
     expect(result.summaryIndex).toEqual(expectedIndex);
@@ -175,66 +203,6 @@ describe('BaseClient', () => {
     expect(result.messagesToRefine).toEqual(expectedMessagesToRefine);
   });
 
-  test('handles context strategy correctly in handleContextStrategy()', async () => {
-    TestClient.addInstructions = jest
-      .fn()
-      .mockReturnValue([
-        { content: 'Hello' },
-        { content: 'How can I help you?' },
-        { content: 'Please provide more details.' },
-        { content: 'I can assist you with that.' },
-      ]);
-    TestClient.getMessagesWithinTokenLimit = jest.fn().mockReturnValue({
-      context: [
-        { content: 'How can I help you?' },
-        { content: 'Please provide more details.' },
-        { content: 'I can assist you with that.' },
-      ],
-      remainingContextTokens: 80,
-      messagesToRefine: [{ content: 'Hello' }],
-      summaryIndex: 3,
-    });
-
-    TestClient.getTokenCount = jest.fn().mockReturnValue(40);
-
-    const instructions = { content: 'Please provide more details.' };
-    const orderedMessages = [
-      { content: 'Hello' },
-      { content: 'How can I help you?' },
-      { content: 'Please provide more details.' },
-      { content: 'I can assist you with that.' },
-    ];
-    const formattedMessages = [
-      { content: 'Hello' },
-      { content: 'How can I help you?' },
-      { content: 'Please provide more details.' },
-      { content: 'I can assist you with that.' },
-    ];
-    const expectedResult = {
-      payload: [
-        {
-          role: 'system',
-          content: 'Refined answer',
-        },
-        { content: 'How can I help you?' },
-        { content: 'Please provide more details.' },
-        { content: 'I can assist you with that.' },
-      ],
-      promptTokens: expect.any(Number),
-      tokenCountMap: {},
-      messages: expect.any(Array),
-    };
-
-    TestClient.shouldSummarize = true;
-    const result = await TestClient.handleContextStrategy({
-      instructions,
-      orderedMessages,
-      formattedMessages,
-    });
-
-    expect(result).toEqual(expectedResult);
-  });
-
   describe('getMessagesForConversation', () => {
     it('should return an empty array if the parentMessageId does not exist', () => {
       const result = TestClient.constructor.getMessagesForConversation({
@@ -293,7 +261,7 @@ describe('BaseClient', () => {
     const unorderedMessages = [
       { id: '3', parentMessageId: '2', text: 'Message 3' },
       { id: '2', parentMessageId: '1', text: 'Message 2' },
-      { id: '1', parentMessageId: '00000000-0000-0000-0000-000000000000', text: 'Message 1' },
+      { id: '1', parentMessageId: Constants.NO_PARENT, text: 'Message 1' },
     ];
 
     it('should return ordered messages based on parentMessageId', () => {
@@ -302,7 +270,7 @@ describe('BaseClient', () => {
         parentMessageId: '3',
       });
       expect(result).toEqual([
-        { id: '1', parentMessageId: '00000000-0000-0000-0000-000000000000', text: 'Message 1' },
+        { id: '1', parentMessageId: Constants.NO_PARENT, text: 'Message 1' },
         { id: '2', parentMessageId: '1', text: 'Message 2' },
         { id: '3', parentMessageId: '2', text: 'Message 3' },
       ]);
@@ -529,9 +497,9 @@ describe('BaseClient', () => {
       );
     });
 
-    test('setOptions is called with the correct arguments', async () => {
+    test('setOptions is called with the correct arguments only when replaceOptions is set to true', async () => {
       TestClient.setOptions = jest.fn();
-      const opts = { conversationId: '123', parentMessageId: '456' };
+      const opts = { conversationId: '123', parentMessageId: '456', replaceOptions: true };
       await TestClient.sendMessage('Hello, world!', opts);
       expect(TestClient.setOptions).toHaveBeenCalledWith(opts);
       TestClient.setOptions.mockClear();
@@ -550,18 +518,24 @@ describe('BaseClient', () => {
       const getReqData = jest.fn();
       const opts = { getReqData };
       const response = await TestClient.sendMessage('Hello, world!', opts);
-      expect(getReqData).toHaveBeenCalledWith({
-        userMessage: expect.objectContaining({ text: 'Hello, world!' }),
-        conversationId: response.conversationId,
-        responseMessageId: response.messageId,
-      });
+      expect(getReqData).toHaveBeenCalledWith(
+        expect.objectContaining({
+          userMessage: expect.objectContaining({ text: 'Hello, world!' }),
+          conversationId: response.conversationId,
+          responseMessageId: response.messageId,
+        }),
+      );
     });
 
     test('onStart is called with the correct arguments', async () => {
       const onStart = jest.fn();
       const opts = { onStart };
       await TestClient.sendMessage('Hello, world!', opts);
-      expect(onStart).toHaveBeenCalledWith(expect.objectContaining({ text: 'Hello, world!' }));
+
+      expect(onStart).toHaveBeenCalledWith(
+        expect.objectContaining({ text: 'Hello, world!' }),
+        expect.any(String),
+      );
     });
 
     test('saveMessageToDatabase is called with the correct arguments', async () => {
@@ -594,9 +568,9 @@ describe('BaseClient', () => {
     test('getTokenCount for response is called with the correct arguments', async () => {
       const tokenCountMap = {}; // Mock tokenCountMap
       TestClient.buildMessages.mockReturnValue({ prompt: [], tokenCountMap });
-      TestClient.getTokenCount = jest.fn();
+      TestClient.getTokenCountForResponse = jest.fn();
       const response = await TestClient.sendMessage('Hello, world!', {});
-      expect(TestClient.getTokenCount).toHaveBeenCalledWith(response.text);
+      expect(TestClient.getTokenCountForResponse).toHaveBeenCalledWith(response);
     });
 
     test('returns an object with the correct shape', async () => {
@@ -612,5 +586,140 @@ describe('BaseClient', () => {
         }),
       );
     });
+
+    test('userMessagePromise is awaited before saving response message', async () => {
+      // Mock the saveMessageToDatabase method
+      TestClient.saveMessageToDatabase = jest.fn().mockImplementation(() => {
+        return new Promise((resolve) => setTimeout(resolve, 100)); // Simulate a delay
+      });
+
+      // Send a message
+      const messagePromise = TestClient.sendMessage('Hello, world!');
+
+      // Wait a short time to ensure the user message save has started
+      await new Promise((resolve) => setTimeout(resolve, 50));
+
+      // Check that saveMessageToDatabase has been called once (for the user message)
+      expect(TestClient.saveMessageToDatabase).toHaveBeenCalledTimes(1);
+
+      // Wait for the message to be fully processed
+      await messagePromise;
+
+      // Check that saveMessageToDatabase has been called twice (once for user message, once for response)
+      expect(TestClient.saveMessageToDatabase).toHaveBeenCalledTimes(2);
+
+      // Check the order of calls
+      const calls = TestClient.saveMessageToDatabase.mock.calls;
+      expect(calls[0][0].isCreatedByUser).toBe(true); // First call should be for user message
+      expect(calls[1][0].isCreatedByUser).toBe(false); // Second call should be for response message
+    });
+  });
+
+  describe('getMessagesWithinTokenLimit with instructions', () => {
+    test('should always include instructions when present', async () => {
+      TestClient.maxContextTokens = 50;
+      const instructions = {
+        role: 'system',
+        content: 'System instructions',
+        tokenCount: 20,
+      };
+
+      const messages = [
+        instructions,
+        { role: 'user', content: 'Hello', tokenCount: 10 },
+        { role: 'assistant', content: 'Hi there', tokenCount: 15 },
+      ];
+
+      const result = await TestClient.getMessagesWithinTokenLimit({
+        messages,
+        instructions,
+      });
+
+      expect(result.context[0]).toBe(instructions);
+      expect(result.remainingContextTokens).toBe(2);
+    });
+
+    test('should handle case when messages exceed limit but instructions must be preserved', async () => {
+      TestClient.maxContextTokens = 30;
+      const instructions = {
+        role: 'system',
+        content: 'System instructions',
+        tokenCount: 20,
+      };
+
+      const messages = [
+        instructions,
+        { role: 'user', content: 'Hello', tokenCount: 10 },
+        { role: 'assistant', content: 'Hi there', tokenCount: 15 },
+      ];
+
+      const result = await TestClient.getMessagesWithinTokenLimit({
+        messages,
+        instructions,
+      });
+
+      // Should only include instructions and the last message that fits
+      expect(result.context).toHaveLength(1);
+      expect(result.context[0].content).toBe(instructions.content);
+      expect(result.messagesToRefine).toHaveLength(2);
+      expect(result.remainingContextTokens).toBe(7); // 30 - 20 - 3 (assistant label)
+    });
+
+    test('should work correctly without instructions (1/2)', async () => {
+      TestClient.maxContextTokens = 50;
+      const messages = [
+        { role: 'user', content: 'Hello', tokenCount: 10 },
+        { role: 'assistant', content: 'Hi there', tokenCount: 15 },
+      ];
+
+      const result = await TestClient.getMessagesWithinTokenLimit({
+        messages,
+      });
+
+      expect(result.context).toHaveLength(2);
+      expect(result.remainingContextTokens).toBe(22); // 50 - 10 - 15 - 3(assistant label)
+      expect(result.messagesToRefine).toHaveLength(0);
+    });
+
+    test('should work correctly without instructions (2/2)', async () => {
+      TestClient.maxContextTokens = 30;
+      const messages = [
+        { role: 'user', content: 'Hello', tokenCount: 10 },
+        { role: 'assistant', content: 'Hi there', tokenCount: 20 },
+      ];
+
+      const result = await TestClient.getMessagesWithinTokenLimit({
+        messages,
+      });
+
+      expect(result.context).toHaveLength(1);
+      expect(result.remainingContextTokens).toBe(7);
+      expect(result.messagesToRefine).toHaveLength(1);
+    });
+
+    test('should handle case when only instructions fit within limit', async () => {
+      TestClient.maxContextTokens = 25;
+      const instructions = {
+        role: 'system',
+        content: 'System instructions',
+        tokenCount: 20,
+      };
+
+      const messages = [
+        instructions,
+        { role: 'user', content: 'Hello', tokenCount: 10 },
+        { role: 'assistant', content: 'Hi there', tokenCount: 15 },
+      ];
+
+      const result = await TestClient.getMessagesWithinTokenLimit({
+        messages,
+        instructions,
+      });
+
+      expect(result.context).toHaveLength(1);
+      expect(result.context[0]).toBe(instructions);
+      expect(result.messagesToRefine).toHaveLength(2);
+      expect(result.remainingContextTokens).toBe(2); // 25 - 20 - 3(assistant label)
+    });
   });
 });

api/app/clients/specs/FakeClient.js

@@ -40,9 +40,9 @@ class FakeClient extends BaseClient {
       };
     }
 
-    this.maxContextTokens = getModelMaxTokens(this.modelOptions.model) ?? 4097;
+    this.maxContextTokens =
+      this.options.maxContextTokens ?? getModelMaxTokens(this.modelOptions.model) ?? 4097;
   }
-  getCompletion() {}
   buildMessages() {}
   getTokenCount(str) {
     return str.length;
@@ -86,6 +86,19 @@ const initializeFakeClient = (apiKey, options, fakeMessages) => {
     return 'Mock response text';
   });
 
+  // eslint-disable-next-line no-unused-vars
+  TestClient.getCompletion = jest.fn().mockImplementation(async (..._args) => {
+    return {
+      choices: [
+        {
+          message: {
+            content: 'Mock response text',
+          },
+        },
+      ],
+    };
+  });
+
   TestClient.buildMessages = jest.fn(async (messages, parentMessageId) => {
     const orderedMessages = TestClient.constructor.getMessagesForConversation({
       messages,

api/app/clients/specs/OpenAIClient.test.js

@@ -1,10 +1,148 @@
+jest.mock('~/cache/getLogStores');
 require('dotenv').config();
+const OpenAI = require('openai');
+const getLogStores = require('~/cache/getLogStores');
+const { fetchEventSource } = require('@waylaidwanderer/fetch-event-source');
+const { genAzureChatCompletion } = require('~/utils/azureUtils');
 const OpenAIClient = require('../OpenAIClient');
-
 jest.mock('meilisearch');
 
+jest.mock('~/lib/db/connectDb');
+jest.mock('~/models', () => ({
+  User: jest.fn(),
+  Key: jest.fn(),
+  Session: jest.fn(),
+  Balance: jest.fn(),
+  Transaction: jest.fn(),
+  getMessages: jest.fn().mockResolvedValue([]),
+  saveMessage: jest.fn(),
+  updateMessage: jest.fn(),
+  deleteMessagesSince: jest.fn(),
+  deleteMessages: jest.fn(),
+  getConvoTitle: jest.fn(),
+  getConvo: jest.fn(),
+  saveConvo: jest.fn(),
+  deleteConvos: jest.fn(),
+  getPreset: jest.fn(),
+  getPresets: jest.fn(),
+  savePreset: jest.fn(),
+  deletePresets: jest.fn(),
+  findFileById: jest.fn(),
+  createFile: jest.fn(),
+  updateFile: jest.fn(),
+  deleteFile: jest.fn(),
+  deleteFiles: jest.fn(),
+  getFiles: jest.fn(),
+  updateFileUsage: jest.fn(),
+}));
+
+jest.mock('@langchain/openai', () => {
+  return {
+    ChatOpenAI: jest.fn().mockImplementation(() => {
+      return {};
+    }),
+  };
+});
+
+jest.mock('openai');
+
+jest.spyOn(OpenAI, 'constructor').mockImplementation(function (...options) {
+  // We can add additional logic here if needed
+  return new OpenAI(...options);
+});
+
+const finalChatCompletion = jest.fn().mockResolvedValue({
+  choices: [
+    {
+      message: { role: 'assistant', content: 'Mock message content' },
+      finish_reason: 'Mock finish reason',
+    },
+  ],
+});
+
+const stream = jest.fn().mockImplementation(() => {
+  let isDone = false;
+  let isError = false;
+  let errorCallback = null;
+
+  const onEventHandlers = {
+    abort: () => {
+      // Mock abort behavior
+    },
+    error: (callback) => {
+      errorCallback = callback; // Save the error callback for later use
+    },
+    finalMessage: (callback) => {
+      callback({ role: 'assistant', content: 'Mock Response' });
+      isDone = true; // Set stream to done
+    },
+  };
+
+  const mockStream = {
+    on: jest.fn((event, callback) => {
+      if (onEventHandlers[event]) {
+        onEventHandlers[event](callback);
+      }
+      return mockStream;
+    }),
+    finalChatCompletion,
+    controller: { abort: jest.fn() },
+    triggerError: () => {
+      isError = true;
+      if (errorCallback) {
+        errorCallback(new Error('Mock error'));
+      }
+    },
+    [Symbol.asyncIterator]: () => {
+      return {
+        next: () => {
+          if (isError) {
+            return Promise.reject(new Error('Mock error'));
+          }
+          if (isDone) {
+            return Promise.resolve({ done: true });
+          }
+          const chunk = { choices: [{ delta: { content: 'Mock chunk' } }] };
+          return Promise.resolve({ value: chunk, done: false });
+        },
+      };
+    },
+  };
+  return mockStream;
+});
+
+const create = jest.fn().mockResolvedValue({
+  choices: [
+    {
+      message: { content: 'Mock message content' },
+      finish_reason: 'Mock finish reason',
+    },
+  ],
+});
+
+OpenAI.mockImplementation(() => ({
+  beta: {
+    chat: {
+      completions: {
+        stream,
+      },
+    },
+  },
+  chat: {
+    completions: {
+      create,
+    },
+  },
+}));
+
 describe('OpenAIClient', () => {
-  let client, client2;
+  const mockSet = jest.fn();
+  const mockCache = { set: mockSet };
+
+  beforeEach(() => {
+    getLogStores.mockReturnValue(mockCache);
+  });
+  let client;
   const model = 'gpt-4';
   const parentMessageId = '1';
   const messages = [
@@ -12,25 +150,40 @@ describe('OpenAIClient', () => {
     { role: 'assistant', sender: 'Assistant', text: 'Hi', messageId: '2' },
   ];
 
+  const defaultOptions = {
+    // debug: true,
+    req: {},
+    openaiApiKey: 'new-api-key',
+    modelOptions: {
+      model,
+      temperature: 0.7,
+    },
+  };
+
+  const defaultAzureOptions = {
+    azureOpenAIApiInstanceName: 'your-instance-name',
+    azureOpenAIApiDeploymentName: 'your-deployment-name',
+    azureOpenAIApiVersion: '2020-07-01-preview',
+  };
+
+  let originalWarn;
+
   beforeAll(() => {
-    jest.spyOn(console, 'warn').mockImplementation(() => {});
+    originalWarn = console.warn;
+    console.warn = jest.fn();
   });
 
   afterAll(() => {
-    console.warn.mockRestore();
+    console.warn = originalWarn;
   });
 
   beforeEach(() => {
-    const options = {
-      // debug: true,
-      openaiApiKey: 'new-api-key',
-      modelOptions: {
-        model,
-        temperature: 0.7,
-      },
-    };
+    console.warn.mockClear();
+  });
+
+  beforeEach(() => {
+    const options = { ...defaultOptions };
     client = new OpenAIClient('test-api-key', options);
-    client2 = new OpenAIClient('test-api-key', options);
     client.summarizeMessages = jest.fn().mockResolvedValue({
       role: 'assistant',
       content: 'Refined answer',
@@ -39,7 +192,7 @@ describe('OpenAIClient', () => {
     client.buildPrompt = jest
       .fn()
       .mockResolvedValue({ prompt: messages.map((m) => m.text).join('\n') });
-    client.constructor.freeAndResetAllEncoders();
+    client.getMessages = jest.fn().mockResolvedValue([]);
   });
 
   describe('setOptions', () => {
@@ -49,14 +202,6 @@ describe('OpenAIClient', () => {
       expect(client.modelOptions.temperature).toBe(0.7);
     });
 
-    it('should set apiKey and useOpenRouter if OPENROUTER_API_KEY is present', () => {
-      process.env.OPENROUTER_API_KEY = 'openrouter-key';
-      client.setOptions({});
-      expect(client.apiKey).toBe('openrouter-key');
-      expect(client.useOpenRouter).toBe(true);
-      delete process.env.OPENROUTER_API_KEY; // Cleanup
-    });
-
     it('should set FORCE_PROMPT based on OPENAI_FORCE_PROMPT or reverseProxyUrl', () => {
       process.env.OPENAI_FORCE_PROMPT = 'true';
       client.setOptions({});
@@ -74,7 +219,7 @@ describe('OpenAIClient', () => {
 
     it('should set isChatCompletion based on useOpenRouter, reverseProxyUrl, or model', () => {
       client.setOptions({ reverseProxyUrl: null });
-      // true by default since default model will be gpt-3.5-turbo
+      // true by default since default model will be gpt-4o-mini
       expect(client.isChatCompletion).toBe(true);
       client.isChatCompletion = undefined;
 
@@ -83,7 +228,7 @@ describe('OpenAIClient', () => {
       expect(client.isChatCompletion).toBe(false);
       client.isChatCompletion = undefined;
 
-      client.setOptions({ modelOptions: { model: 'gpt-3.5-turbo' }, reverseProxyUrl: null });
+      client.setOptions({ modelOptions: { model: 'gpt-4o-mini' }, reverseProxyUrl: null });
       expect(client.isChatCompletion).toBe(true);
     });
 
@@ -94,7 +239,7 @@ describe('OpenAIClient', () => {
 
       client.setOptions({ reverseProxyUrl: 'https://example.com/completions' });
       expect(client.completionsUrl).toBe('https://example.com/completions');
-      expect(client.langchainProxy).toBe(null);
+      expect(client.langchainProxy).toBe('https://example.com/completions');
     });
   });
 
@@ -188,83 +333,18 @@ describe('OpenAIClient', () => {
     });
   });
 
-  describe('selectTokenizer', () => {
-    it('should get the correct tokenizer based on the instance state', () => {
-      const tokenizer = client.selectTokenizer();
-      expect(tokenizer).toBeDefined();
-    });
-  });
-
-  describe('freeAllTokenizers', () => {
-    it('should free all tokenizers', () => {
-      // Create a tokenizer
-      const tokenizer = client.selectTokenizer();
-
-      // Mock 'free' method on the tokenizer
-      tokenizer.free = jest.fn();
-
-      client.constructor.freeAndResetAllEncoders();
-
-      // Check if 'free' method has been called on the tokenizer
-      expect(tokenizer.free).toHaveBeenCalled();
-    });
-  });
-
   describe('getTokenCount', () => {
     it('should return the correct token count', () => {
       const count = client.getTokenCount('Hello, world!');
       expect(count).toBeGreaterThan(0);
     });
-
-    it('should reset the encoder and count when count reaches 25', () => {
-      const freeAndResetEncoderSpy = jest.spyOn(client.constructor, 'freeAndResetAllEncoders');
-
-      // Call getTokenCount 25 times
-      for (let i = 0; i < 25; i++) {
-        client.getTokenCount('test text');
-      }
-
-      expect(freeAndResetEncoderSpy).toHaveBeenCalled();
-    });
-
-    it('should not reset the encoder and count when count is less than 25', () => {
-      const freeAndResetEncoderSpy = jest.spyOn(client.constructor, 'freeAndResetAllEncoders');
-      freeAndResetEncoderSpy.mockClear();
-
-      // Call getTokenCount 24 times
-      for (let i = 0; i < 24; i++) {
-        client.getTokenCount('test text');
-      }
-
-      expect(freeAndResetEncoderSpy).not.toHaveBeenCalled();
-    });
-
-    it('should handle errors and reset the encoder', () => {
-      const freeAndResetEncoderSpy = jest.spyOn(client.constructor, 'freeAndResetAllEncoders');
-
-      // Mock encode function to throw an error
-      client.selectTokenizer().encode = jest.fn().mockImplementation(() => {
-        throw new Error('Test error');
-      });
-
-      client.getTokenCount('test text');
-
-      expect(freeAndResetEncoderSpy).toHaveBeenCalled();
-    });
-
-    it('should not throw null pointer error when freeing the same encoder twice', () => {
-      client.constructor.freeAndResetAllEncoders();
-      client2.constructor.freeAndResetAllEncoders();
-
-      const count = client2.getTokenCount('test text');
-      expect(count).toBeGreaterThan(0);
-    });
   });
 
   describe('getSaveOptions', () => {
     it('should return the correct save options', () => {
       const options = client.getSaveOptions();
       expect(options).toHaveProperty('chatGptLabel');
+      expect(options).toHaveProperty('modelLabel');
       expect(options).toHaveProperty('promptPrefix');
     });
   });
@@ -299,7 +379,7 @@ describe('OpenAIClient', () => {
         promptPrefix: 'Test Prefix',
       });
       expect(result).toHaveProperty('prompt');
-      const instructions = result.prompt.find((item) => item.name === 'instructions');
+      const instructions = result.prompt.find((item) => item.content.includes('Test Prefix'));
       expect(instructions).toBeDefined();
       expect(instructions.content).toContain('Test Prefix');
     });
@@ -329,7 +409,9 @@ describe('OpenAIClient', () => {
       const result = await client.buildMessages(messages, parentMessageId, {
         isChatCompletion: true,
       });
-      const instructions = result.prompt.find((item) => item.name === 'instructions');
+      const instructions = result.prompt.find((item) =>
+        item.content.includes('Test Prefix from options'),
+      );
       expect(instructions.content).toContain('Test Prefix from options');
     });
 
@@ -337,7 +419,7 @@ describe('OpenAIClient', () => {
       const result = await client.buildMessages(messages, parentMessageId, {
         isChatCompletion: true,
       });
-      const instructions = result.prompt.find((item) => item.name === 'instructions');
+      const instructions = result.prompt.find((item) => item.content.includes('Test Prefix'));
       expect(instructions).toBeUndefined();
     });
 
@@ -398,7 +480,6 @@ describe('OpenAIClient', () => {
     testCases.forEach((testCase) => {
       it(`should return ${testCase.expected} tokens for model ${testCase.model}`, () => {
         client.modelOptions.model = testCase.model;
-        client.selectTokenizer();
         // 3 tokens for assistant label
         let totalTokens = 3;
         for (let message of example_messages) {
@@ -407,5 +488,207 @@ describe('OpenAIClient', () => {
         expect(totalTokens).toBe(testCase.expected);
       });
     });
+
+    const vision_request = [
+      {
+        role: 'user',
+        content: [
+          {
+            type: 'text',
+            text: 'describe what is in this image?',
+          },
+          {
+            type: 'image_url',
+            image_url: {
+              url: 'https://venturebeat.com/wp-content/uploads/2019/03/openai-1.png',
+              detail: 'high',
+            },
+          },
+        ],
+      },
+    ];
+
+    const expectedTokens = 14;
+    const visionModel = 'gpt-4-vision-preview';
+
+    it(`should return ${expectedTokens} tokens for model ${visionModel} (Vision Request)`, () => {
+      client.modelOptions.model = visionModel;
+      // 3 tokens for assistant label
+      let totalTokens = 3;
+      for (let message of vision_request) {
+        totalTokens += client.getTokenCountForMessage(message);
+      }
+      expect(totalTokens).toBe(expectedTokens);
+    });
+  });
+
+  describe('sendMessage/getCompletion/chatCompletion', () => {
+    afterEach(() => {
+      delete process.env.AZURE_OPENAI_DEFAULT_MODEL;
+      delete process.env.AZURE_USE_MODEL_AS_DEPLOYMENT_NAME;
+    });
+
+    it('should call getCompletion and fetchEventSource when using a text/instruct model', async () => {
+      const model = 'text-davinci-003';
+      const onProgress = jest.fn().mockImplementation(() => ({}));
+
+      const testClient = new OpenAIClient('test-api-key', {
+        ...defaultOptions,
+        modelOptions: { model },
+      });
+
+      const getCompletion = jest.spyOn(testClient, 'getCompletion');
+      await testClient.sendMessage('Hi mom!', { onProgress });
+
+      expect(getCompletion).toHaveBeenCalled();
+      expect(getCompletion.mock.calls.length).toBe(1);
+
+      expect(getCompletion.mock.calls[0][0]).toBe('||>User:\nHi mom!\n||>Assistant:\n');
+
+      expect(fetchEventSource).toHaveBeenCalled();
+      expect(fetchEventSource.mock.calls.length).toBe(1);
+
+      // Check if the first argument (url) is correct
+      const firstCallArgs = fetchEventSource.mock.calls[0];
+
+      const expectedURL = 'https://api.openai.com/v1/completions';
+      expect(firstCallArgs[0]).toBe(expectedURL);
+
+      const requestBody = JSON.parse(firstCallArgs[1].body);
+      expect(requestBody).toHaveProperty('model');
+      expect(requestBody.model).toBe(model);
+    });
+
+    it('[Azure OpenAI] should call chatCompletion and OpenAI.stream with correct args', async () => {
+      // Set a default model
+      process.env.AZURE_OPENAI_DEFAULT_MODEL = 'gpt4-turbo';
+
+      const onProgress = jest.fn().mockImplementation(() => ({}));
+      client.azure = defaultAzureOptions;
+      const chatCompletion = jest.spyOn(client, 'chatCompletion');
+      await client.sendMessage('Hi mom!', {
+        replaceOptions: true,
+        ...defaultOptions,
+        modelOptions: { model: 'gpt4-turbo', stream: true },
+        onProgress,
+        azure: defaultAzureOptions,
+      });
+
+      expect(chatCompletion).toHaveBeenCalled();
+      expect(chatCompletion.mock.calls.length).toBe(1);
+
+      const chatCompletionArgs = chatCompletion.mock.calls[0][0];
+      const { payload } = chatCompletionArgs;
+
+      expect(payload[0].role).toBe('user');
+      expect(payload[0].content).toBe('Hi mom!');
+
+      // Azure OpenAI does not use the model property, and will error if it's passed
+      // This check ensures the model property is not present
+      const streamArgs = stream.mock.calls[0][0];
+      expect(streamArgs).not.toHaveProperty('model');
+
+      // Check if the baseURL is correct
+      const constructorArgs = OpenAI.mock.calls[0][0];
+      const expectedURL = genAzureChatCompletion(defaultAzureOptions).split('/chat')[0];
+      expect(constructorArgs.baseURL).toBe(expectedURL);
+    });
+  });
+
+  describe('checkVisionRequest functionality', () => {
+    let client;
+    const attachments = [{ type: 'image/png' }];
+
+    beforeEach(() => {
+      client = new OpenAIClient('test-api-key', {
+        endpoint: 'ollama',
+        modelOptions: {
+          model: 'initial-model',
+        },
+        modelsConfig: {
+          ollama: ['initial-model', 'llava', 'other-model'],
+        },
+      });
+
+      client.defaultVisionModel = 'non-valid-default-model';
+    });
+
+    afterEach(() => {
+      jest.restoreAllMocks();
+    });
+
+    it('should set "llava" as the model if it is the first valid model when default validation fails', () => {
+      client.checkVisionRequest(attachments);
+
+      expect(client.modelOptions.model).toBe('llava');
+      expect(client.isVisionModel).toBeTruthy();
+      expect(client.modelOptions.stop).toBeUndefined();
+    });
+  });
+
+  describe('getStreamUsage', () => {
+    it('should return this.usage when completion_tokens_details is null', () => {
+      const client = new OpenAIClient('test-api-key', defaultOptions);
+      client.usage = {
+        completion_tokens_details: null,
+        prompt_tokens: 10,
+        completion_tokens: 20,
+      };
+      client.inputTokensKey = 'prompt_tokens';
+      client.outputTokensKey = 'completion_tokens';
+
+      const result = client.getStreamUsage();
+
+      expect(result).toEqual(client.usage);
+    });
+
+    it('should return this.usage when completion_tokens_details is missing reasoning_tokens', () => {
+      const client = new OpenAIClient('test-api-key', defaultOptions);
+      client.usage = {
+        completion_tokens_details: {
+          other_tokens: 5,
+        },
+        prompt_tokens: 10,
+        completion_tokens: 20,
+      };
+      client.inputTokensKey = 'prompt_tokens';
+      client.outputTokensKey = 'completion_tokens';
+
+      const result = client.getStreamUsage();
+
+      expect(result).toEqual(client.usage);
+    });
+
+    it('should calculate output tokens correctly when completion_tokens_details is present with reasoning_tokens', () => {
+      const client = new OpenAIClient('test-api-key', defaultOptions);
+      client.usage = {
+        completion_tokens_details: {
+          reasoning_tokens: 30,
+          other_tokens: 5,
+        },
+        prompt_tokens: 10,
+        completion_tokens: 20,
+      };
+      client.inputTokensKey = 'prompt_tokens';
+      client.outputTokensKey = 'completion_tokens';
+
+      const result = client.getStreamUsage();
+
+      expect(result).toEqual({
+        reasoning_tokens: 30,
+        other_tokens: 5,
+        prompt_tokens: 10,
+        completion_tokens: 10, // |30 - 20| = 10
+      });
+    });
+
+    it('should return this.usage when it is undefined', () => {
+      const client = new OpenAIClient('test-api-key', defaultOptions);
+      client.usage = undefined;
+
+      const result = client.getStreamUsage();
+
+      expect(result).toBeUndefined();
+    });
   });
 });

api/app/clients/specs/OpenAIClient.tokens.js

@@ -38,7 +38,12 @@ const run = async () => {
   "On the other hand, we denounce with righteous indignation and dislike men who are so beguiled and demoralized by the charms of pleasure of the moment, so blinded by desire, that they cannot foresee the pain and trouble that are bound to ensue; and equal blame belongs to those who fail in their duty through weakness of will, which is the same as saying through shrinking from toil and pain. These cases are perfectly simple and easy to distinguish. In a free hour, when our power of choice is untrammelled and when nothing prevents our being able to do what we like best, every pleasure is to be welcomed and every pain avoided. But in certain circumstances and owing to the claims of duty or the obligations of business it will frequently occur that pleasures have to be repudiated and annoyances accepted. The wise man therefore always holds in these matters to this principle of selection: he rejects pleasures to secure other greater pleasures, or else he endures pains to avoid worse pains."
   `;
   const model = 'gpt-3.5-turbo';
-  const maxContextTokens = model === 'gpt-4' ? 8191 : model === 'gpt-4-32k' ? 32767 : 4095; // 1 less than maximum
+  let maxContextTokens = 4095;
+  if (model === 'gpt-4') {
+    maxContextTokens = 8191;
+  } else if (model === 'gpt-4-32k') {
+    maxContextTokens = 32767;
+  }
   const clientOptions = {
     reverseProxyUrl: process.env.OPENAI_REVERSE_PROXY || null,
     maxContextTokens,

api/app/clients/specs/PluginsClient.test.js

@@ -1,9 +1,10 @@
-const { HumanChatMessage, AIChatMessage } = require('langchain/schema');
-const PluginsClient = require('../PluginsClient');
 const crypto = require('crypto');
+const { Constants } = require('librechat-data-provider');
+const { HumanMessage, AIMessage } = require('@langchain/core/messages');
+const PluginsClient = require('../PluginsClient');
 
-jest.mock('../../../lib/db/connectDb');
-jest.mock('../../../models/Conversation', () => {
+jest.mock('~/lib/db/connectDb');
+jest.mock('~/models/Conversation', () => {
   return function () {
     return {
       save: jest.fn(),
@@ -12,6 +13,12 @@ jest.mock('../../../models/Conversation', () => {
   };
 });
 
+const defaultAzureOptions = {
+  azureOpenAIApiInstanceName: 'your-instance-name',
+  azureOpenAIApiDeploymentName: 'your-deployment-name',
+  azureOpenAIApiVersion: '2020-07-01-preview',
+};
+
 describe('PluginsClient', () => {
   let TestAgent;
   let options = {
@@ -48,8 +55,8 @@ describe('PluginsClient', () => {
 
         const chatMessages = orderedMessages.map((msg) =>
           msg?.isCreatedByUser || msg?.role?.toLowerCase() === 'user'
-            ? new HumanChatMessage(msg.text)
-            : new AIChatMessage(msg.text),
+            ? new HumanMessage(msg.text)
+            : new AIMessage(msg.text),
         );
 
         TestAgent.currentMessages = orderedMessages;
@@ -60,7 +67,7 @@ describe('PluginsClient', () => {
         TestAgent.setOptions(opts);
       }
       const conversationId = opts.conversationId || crypto.randomUUID();
-      const parentMessageId = opts.parentMessageId || '00000000-0000-0000-0000-000000000000';
+      const parentMessageId = opts.parentMessageId || Constants.NO_PARENT;
       const userMessageId = opts.overrideParentMessageId || crypto.randomUUID();
       this.pastMessages = await TestAgent.loadHistory(
         conversationId,
@@ -144,4 +151,164 @@ describe('PluginsClient', () => {
       expect(chatMessages[0].text).toEqual(userMessage);
     });
   });
+
+  describe('getFunctionModelName', () => {
+    let client;
+
+    beforeEach(() => {
+      client = new PluginsClient('dummy_api_key');
+    });
+
+    test('should return the input when it includes a dash followed by four digits', () => {
+      expect(client.getFunctionModelName('-1234')).toBe('-1234');
+      expect(client.getFunctionModelName('gpt-4-5678-preview')).toBe('gpt-4-5678-preview');
+    });
+
+    test('should return the input for all function-capable models (`0613` models and above)', () => {
+      expect(client.getFunctionModelName('gpt-4-0613')).toBe('gpt-4-0613');
+      expect(client.getFunctionModelName('gpt-4-32k-0613')).toBe('gpt-4-32k-0613');
+      expect(client.getFunctionModelName('gpt-3.5-turbo-0613')).toBe('gpt-3.5-turbo-0613');
+      expect(client.getFunctionModelName('gpt-3.5-turbo-16k-0613')).toBe('gpt-3.5-turbo-16k-0613');
+      expect(client.getFunctionModelName('gpt-3.5-turbo-1106')).toBe('gpt-3.5-turbo-1106');
+      expect(client.getFunctionModelName('gpt-4-1106-preview')).toBe('gpt-4-1106-preview');
+      expect(client.getFunctionModelName('gpt-4-1106')).toBe('gpt-4-1106');
+    });
+
+    test('should return the corresponding model if input is non-function capable (`0314` models)', () => {
+      expect(client.getFunctionModelName('gpt-4-0314')).toBe('gpt-4');
+      expect(client.getFunctionModelName('gpt-4-32k-0314')).toBe('gpt-4');
+      expect(client.getFunctionModelName('gpt-3.5-turbo-0314')).toBe('gpt-3.5-turbo');
+      expect(client.getFunctionModelName('gpt-3.5-turbo-16k-0314')).toBe('gpt-3.5-turbo');
+    });
+
+    test('should return "gpt-3.5-turbo" when the input includes "gpt-3.5-turbo"', () => {
+      expect(client.getFunctionModelName('test gpt-3.5-turbo model')).toBe('gpt-3.5-turbo');
+    });
+
+    test('should return "gpt-4" when the input includes "gpt-4"', () => {
+      expect(client.getFunctionModelName('testing gpt-4')).toBe('gpt-4');
+    });
+
+    test('should return "gpt-3.5-turbo" for input that does not meet any specific condition', () => {
+      expect(client.getFunctionModelName('random string')).toBe('gpt-3.5-turbo');
+      expect(client.getFunctionModelName('')).toBe('gpt-3.5-turbo');
+    });
+  });
+
+  describe('Azure OpenAI tests specific to Plugins', () => {
+    // TODO: add more tests for Azure OpenAI integration with Plugins
+    // let client;
+    // beforeEach(() => {
+    //   client = new PluginsClient('dummy_api_key');
+    // });
+
+    test('should not call getFunctionModelName when azure options are set', () => {
+      const spy = jest.spyOn(PluginsClient.prototype, 'getFunctionModelName');
+      const model = 'gpt-4-turbo';
+
+      // note, without the azure change in PR #1766, `getFunctionModelName` is called twice
+      const testClient = new PluginsClient('dummy_api_key', {
+        agentOptions: {
+          model,
+          agent: 'functions',
+        },
+        azure: defaultAzureOptions,
+      });
+
+      expect(spy).not.toHaveBeenCalled();
+      expect(testClient.agentOptions.model).toBe(model);
+
+      spy.mockRestore();
+    });
+  });
+
+  describe('sendMessage with filtered tools', () => {
+    let TestAgent;
+    const apiKey = 'fake-api-key';
+    const mockTools = [{ name: 'tool1' }, { name: 'tool2' }, { name: 'tool3' }, { name: 'tool4' }];
+
+    beforeEach(() => {
+      TestAgent = new PluginsClient(apiKey, {
+        tools: mockTools,
+        modelOptions: {
+          model: 'gpt-3.5-turbo',
+          temperature: 0,
+          max_tokens: 2,
+        },
+        agentOptions: {
+          model: 'gpt-3.5-turbo',
+        },
+      });
+
+      TestAgent.options.req = {
+        app: {
+          locals: {},
+        },
+      };
+
+      TestAgent.sendMessage = jest.fn().mockImplementation(async () => {
+        const { filteredTools = [], includedTools = [] } = TestAgent.options.req.app.locals;
+
+        if (includedTools.length > 0) {
+          const tools = TestAgent.options.tools.filter((plugin) =>
+            includedTools.includes(plugin.name),
+          );
+          TestAgent.options.tools = tools;
+        } else {
+          const tools = TestAgent.options.tools.filter(
+            (plugin) => !filteredTools.includes(plugin.name),
+          );
+          TestAgent.options.tools = tools;
+        }
+
+        return {
+          text: 'Mocked response',
+          tools: TestAgent.options.tools,
+        };
+      });
+    });
+
+    test('should filter out tools when filteredTools is provided', async () => {
+      TestAgent.options.req.app.locals.filteredTools = ['tool1', 'tool3'];
+      const response = await TestAgent.sendMessage('Test message');
+      expect(response.tools).toHaveLength(2);
+      expect(response.tools).toEqual(
+        expect.arrayContaining([
+          expect.objectContaining({ name: 'tool2' }),
+          expect.objectContaining({ name: 'tool4' }),
+        ]),
+      );
+    });
+
+    test('should only include specified tools when includedTools is provided', async () => {
+      TestAgent.options.req.app.locals.includedTools = ['tool2', 'tool4'];
+      const response = await TestAgent.sendMessage('Test message');
+      expect(response.tools).toHaveLength(2);
+      expect(response.tools).toEqual(
+        expect.arrayContaining([
+          expect.objectContaining({ name: 'tool2' }),
+          expect.objectContaining({ name: 'tool4' }),
+        ]),
+      );
+    });
+
+    test('should prioritize includedTools over filteredTools', async () => {
+      TestAgent.options.req.app.locals.filteredTools = ['tool1', 'tool3'];
+      TestAgent.options.req.app.locals.includedTools = ['tool1', 'tool2'];
+      const response = await TestAgent.sendMessage('Test message');
+      expect(response.tools).toHaveLength(2);
+      expect(response.tools).toEqual(
+        expect.arrayContaining([
+          expect.objectContaining({ name: 'tool1' }),
+          expect.objectContaining({ name: 'tool2' }),
+        ]),
+      );
+    });
+
+    test('should not modify tools when no filters are provided', async () => {
+      const response = await TestAgent.sendMessage('Test message');
+      expect(response.tools).toHaveLength(4);
+      expect(response.tools).toEqual(expect.arrayContaining(mockTools));
+    });
+  });
 });