LibreChat Integration¶

Connect LibreChat to Amazon Bedrock models through stdapi.ai. Deploy a self-hosted ChatGPT alternative with multi-user support, conversation management, and access to Claude, Amazon Nova, and other Bedrock models.

About LibreChat¶

🔗 Links: Website | GitHub | Documentation | Discord

LibreChat is a feature-rich, open-source AI chat platform with:

• 30,000+ GitHub stars - Popular open-source ChatGPT alternative
• Production-ready - Used by organizations worldwide
• Multi-modal - Chat, voice, images, and document analysis
• Extensible - Plugin system, custom endpoints, and integrations

Why LibreChat + stdapi.ai?¶

Key Benefits:

• Multi-user platform - Team collaboration with individual accounts and conversations
• Privacy - Your conversations and data stay in your infrastructure
• Enterprise models - Access Claude, Amazon Nova, and all Bedrock models
• Rich features - Conversation history, presets, plugins, file uploads
• Cost control - AWS pricing with usage tracking and quotas
• Customizable - White-label, branding, and custom configurations
• Active development - Regular updates with new features

Prerequisites¶

🚀 Quick Start with Docker¶

Step 1: Clone LibreChat Repository¶

Get the latest version of LibreChat from GitHub.

# Clone the repository
git clone https://github.com/danny-avila/LibreChat.git
cd LibreChat

# Copy the example environment file
cp .env.example .env

Step 2: Configure stdapi.ai Connection¶

Edit the .env file to configure LibreChat to use stdapi.ai as the OpenAI provider.

# Core LibreChat settings
HOST=0.0.0.0
PORT=3080

# MongoDB (required for conversation storage)
MONGO_URI=mongodb://mongodb:27017/LibreChat

# Session secret (generate a random string)
SESSION_SECRET=your_random_session_secret_here

# stdapi.ai as OpenAI provider
OPENAI_API_KEY=your_stdapi_key_here
OPENAI_REVERSE_PROXY=https://YOUR_SERVER_URL/v1

# Optional: Enable all features
DEBUG_OPENAI=true

openssl rand -base64 32

Step 3: Configure Available Models¶

Create a librechat.yaml configuration file to define which Bedrock models are available to users.

version: 1.0.5
cache: true

endpoints:
  custom:
    - name: "Amazon Bedrock"
      apiKey: "${OPENAI_API_KEY}"
      baseURL: "https://YOUR_SERVER_URL/v1"
      models:
        default:
          - "anthropic.claude-sonnet-4-5-20250929-v1:0"
          - "anthropic.claude-3-5-haiku-20241022-v1:0"
          - "amazon.nova-pro-v1:0"
          - "amazon.nova-lite-v1:0"
          - "amazon.nova-micro-v1:0"
        fetch: false
      titleConvo: true
      titleModel: "amazon.nova-micro-v1:0"
      summarize: false
      summaryModel: "amazon.nova-lite-v1:0"
      forcePrompt: false
      modelDisplayLabel: "Amazon Bedrock"

Available Models:

All Amazon Bedrock chat models are compatible. Add any model IDs available in your AWS region:

• Anthropic Claude — All Claude model variants (Sonnet, Haiku, Opus)
• Amazon Nova — All Nova family models (Pro, Lite, Micro)
• Meta Llama — Llama models (if enabled)
• Mistral AI — Mistral and Mixtral models (if enabled)
• And more — Any Bedrock chat model

Step 4: Enable Additional Features¶

Enhance your LibreChat instance with file uploads, speech-to-text, and image generation.

# Core settings (from Step 2)
OPENAI_API_KEY=your_stdapi_key_here
OPENAI_REVERSE_PROXY=https://YOUR_SERVER_URL/v1

# File upload and parsing
ENABLE_FILE_UPLOADS=true
FILE_UPLOAD_SIZE_LIMIT=20  # MB

# Speech to Text (Whisper API compatible)
STT_API_KEY=your_stdapi_key_here
STT_API_URL=https://YOUR_SERVER_URL/v1

# Text to Speech
TTS_API_KEY=your_stdapi_key_here
TTS_API_URL=https://YOUR_SERVER_URL/v1

# Image generation
DALLE_API_KEY=your_stdapi_key_here
DALLE_REVERSE_PROXY=https://YOUR_SERVER_URL/v1

# Search (optional, for web search integration)
# SEARCH_API_KEY=your_search_api_key

Step 5: Deploy with Docker Compose¶

Start LibreChat with all dependencies using Docker Compose.

version: '3.8'

services:
  api:
    volumes:
      - ./librechat.yaml:/app/librechat.yaml:ro
    environment:
      - OPENAI_API_KEY=${OPENAI_API_KEY}
      - OPENAI_REVERSE_PROXY=${OPENAI_REVERSE_PROXY}
      - STT_API_KEY=${STT_API_KEY}
      - STT_API_URL=${STT_API_URL}
      - TTS_API_KEY=${TTS_API_KEY}
      - TTS_API_URL=${TTS_API_URL}
      - DALLE_API_KEY=${DALLE_API_KEY}
      - DALLE_REVERSE_PROXY=${DALLE_REVERSE_PROXY}

Start the application:

# Start all services
docker-compose up -d

# View logs
docker-compose logs -f

# Access LibreChat at http://localhost:3080

🎯 What You Can Do Now¶

Once deployed, LibreChat with stdapi.ai provides a full-featured AI assistant platform:

💬 Conversational AI¶

• Multi-Turn Conversations: Contextual discussions with full conversation history
• Model Switching: Change models mid-conversation based on task complexity
• Conversation Management: Save, search, share, and organize conversations
• Presets: Create reusable conversation templates with custom instructions

🎨 Multi-Modal Capabilities¶

• Image Generation: Create images using Amazon Nova Canvas directly in chat
• Voice Input: Speak your messages using speech-to-text (Amazon Transcribe)
• Voice Output: Listen to responses with text-to-speech (Amazon Polly)
• File Analysis: Upload documents for AI analysis and Q&A

👥 Team Collaboration¶

• Multi-User Support: Individual accounts with separate conversation histories
• User Management: Admin controls for user registration and permissions
• Shared Conversations: Export and share conversations with team members
• Usage Tracking: Monitor token usage and costs per user

⚙️ Customization¶

• Custom Prompts: Define system prompts and conversation starters
• Branding: Customize logo, colors, and application name
• Model Presets: Pre-configure optimal settings for specific use cases
• Plugins: Extend functionality with community plugins

🔧 Advanced Configuration¶

User Registration and Authentication¶

Control who can access your LibreChat instance.

# Allow registration (set to false for invite-only)
ALLOW_REGISTRATION=true

# Email verification (requires SMTP)
ALLOW_EMAIL_LOGIN=true
EMAIL_SERVICE=gmail
EMAIL_USERNAME=your-email@gmail.com
EMAIL_PASSWORD=your-app-password
EMAIL_FROM=noreply@yourdomain.com

# Social login (optional)
GOOGLE_CLIENT_ID=your_google_client_id
GOOGLE_CLIENT_SECRET=your_google_client_secret

RAG and Document Intelligence¶

Enable document uploads and semantic search for knowledge base functionality.

endpoints:
  custom:
    - name: "Amazon Bedrock"
      apiKey: "${OPENAI_API_KEY}"
      baseURL: "https://YOUR_SERVER_URL/v1"
      models:
        default:
          - "anthropic.claude-sonnet-4-5-20250929-v1:0"
          - "amazon.nova-pro-v1:0"
          - "amazon.nova-lite-v1:0"
      titleConvo: true
      titleModel: "amazon.nova-micro-v1:0"
      # Enable file uploads for RAG
      fileConfig:
        endpoints:
          - "custom"
        fileLimit: 10
        fileSizeLimit: 20  # MB
        totalSizeLimit: 100  # MB
        supportedMimeTypes:
          - "application/pdf"
          - "text/plain"
          - "text/markdown"
          - "application/vnd.openxmlformats-officedocument.wordprocessingml.document"

Embedding Configuration (.env):

# Use stdapi.ai for embeddings
EMBEDDINGS_PROVIDER=openai
OPENAI_API_KEY=your_stdapi_key_here
EMBEDDINGS_MODEL=amazon.titan-embed-text-v2:0

Production Deployment¶

Deploy LibreChat in production with SSL, persistent storage, and backups.

version: '3.8'

services:
  nginx:
    image: nginx:alpine
    ports:
      - "80:80"
      - "443:443"
    volumes:
      - ./nginx.conf:/etc/nginx/nginx.conf:ro
      - ./ssl:/etc/nginx/ssl:ro
    depends_on:
      - api

  api:
    image: ghcr.io/danny-avila/librechat:latest
    env_file:
      - .env
    volumes:
      - ./librechat.yaml:/app/librechat.yaml:ro
      - librechat-images:/app/client/public/images
      - librechat-logs:/app/api/logs
    depends_on:
      - mongodb
    restart: unless-stopped

  mongodb:
    image: mongo:6.0
    volumes:
      - mongodb-data:/data/db
    restart: unless-stopped
    command: mongod --quiet --logpath /dev/null

volumes:
  mongodb-data:
  librechat-images:
  librechat-logs:

📊 Model Recommendations¶

Choose the right models for different use cases to optimize performance and cost. These are examples—all Bedrock models are available.

💡 Pro Tips & Best Practices¶

🚀 Next Steps & Resources¶

Getting Started¶

1. Access LibreChat: Open http://localhost:3080 and create an account
2. Test Models: Try different models to understand their strengths
3. Configure Presets: Set up conversation templates for your workflows
4. Invite Team: Share access with your team members
5. Monitor Usage: Track token consumption and adjust as needed

Learn More¶

Community & Support¶

⚠️ Important Considerations¶