LibreChat

LibreChat is an open-source AI chat platform that you can self-host. You can configure it to use AI Gateway for unified model access and spend monitoring.

1. Create an API key

   Go to the AI Gateway tab of the Vercel dashboard and click API keys to create a new API key.

2. Install LibreChat

   Clone the LibreChat repository and set up the environment:

   Terminal

   git clone https://github.com/danny-avila/LibreChat.git
   cd LibreChat
   cp .env.example .env

   Windows users: Replace cp with copy if needed. Docker Desktop is required for this setup.

3. Create Docker override file

   Create a docker-compose.override.yml file in your LibreChat root directory to mount the configuration:

   docker-compose.override.yml

   services:
     api:
       volumes:
         - type: bind
           source: ./librechat.yaml
           target: /app/librechat.yaml

   This allows LibreChat to read your custom endpoint configuration.

4. Add API key to environment

   Add your AI Gateway API key to your .env file in the LibreChat root directory:

   .env

   AI_GATEWAY_API_KEY=your-ai-gateway-api-key

   Use the ${"${VARIABLE_NAME}"} pattern to reference environment variables. Do not include raw API keys in the YAML file.

5. Configure custom endpoint

   Create a librechat.yaml file in your LibreChat root directory:

   librechat.yaml

   version: 1.2.8
   cache: true
    
   endpoints:
     custom:
       - name: "Vercel"
         apiKey: "${AI_GATEWAY_API_KEY}"
         baseURL: "https://ai-gateway.vercel.sh/v1"
         titleConvo: true
         models:
           default:
             - "openai/gpt-5.2"
             - "anthropic/claude-sonnet-4.5"
             - "google/gemini-3-flash"
           fetch: true
         titleModel: "openai/gpt-5.2"

   Setting fetch: true automatically fetches all available models from AI Gateway. Browse the full catalog on the models page.

6. Start LibreChat

   Start or restart your LibreChat instance to apply the configuration:

   Terminal

   docker compose up -d

   If LibreChat is already running, restart it:

   Terminal

   docker compose restart

   Once started, navigate to http://localhost:3080/ to access LibreChat.

7. Select AI Gateway endpoint

   In the LibreChat interface:

   1. Click the endpoint dropdown at the top
   2. Select Vercel
   3. Choose a model from the available options

   Your requests will now be routed through AI Gateway. You can verify this by checking your AI Gateway Overview in the Vercel dashboard.

8. (Optional) Monitor usage and spend

   View your usage, spend, and request activity in the AI Gateway tab of the Vercel dashboard. See the observability documentation for more details.

You can customize the LibreChat endpoint configuration:

• titleConvo: Set to true to enable automatic conversation titles
• titleModel: Specify which model to use for generating conversation titles
• modelDisplayLabel: Customize the label shown in the interface (optional)
• dropParams: Remove default parameters that some providers don't support

See the LibreChat custom endpoints documentation for all available options.