wdmtech/openrouter-search-mcp

OpenRouter Search MCP Server

A Model Context Protocol (MCP) server that provides web search functionality using OpenRouter's online-enabled models.

Features

• Web Search: Search the web using OpenRouter's online models
• Customizable Models: Configure which OpenRouter model to use via environment variables or per-request parameters
• Flexible Configuration: Support for both global and per-request model selection

Installation

1. Clone or create this server in your MCP servers directory
2. Install dependencies:

   npm install
   

3. Build the server:

   npm run build
   

Configuration

Environment Variables

• OPENROUTER_API_KEY (required): Your OpenRouter API key
• DEFAULT_MODEL (optional): Default model to use (defaults to openai/gpt-4o:online)

MCP Settings

Add to your MCP settings configuration:

{
  "mcpServers": {
    "openrouter-search": {
      "command": "node",
      "args": ["/path/to/openrouter-search/build/index.js"],
      "env": {
        "OPENROUTER_API_KEY": "your-api-key-here",
        "DEFAULT_MODEL": "openai/gpt-4o:online"
      },
      "disabled": false,
      "autoApprove": []
    }
  }
}

Available Tools

web_search

Search the web using OpenRouter's online models.

Parameters:

• query (string, required): The search query
• model (string, optional): OpenRouter model to use for this search

Example:

{
  "query": "latest developments in AI",
  "model": "anthropic/claude-3-5-sonnet:online"
}

Supported Models

For web search functionality, you can use any OpenRouter model with the :online suffix:

• openai/gpt-4o:online (default)
• anthropic/claude-3-5-sonnet:online
• anthropic/claude-3-haiku:online
• google/gemini-pro:online
• openrouter/auto (automatically selects best model with web plugin)

How it works: Adding :online to any model slug enables web search via OpenRouter's web plugin powered by Exa. This is equivalent to using the web plugin explicitly.

Alternative: You can also use openrouter/auto which automatically selects the best model and enables web search.

Usage Examples

Once configured, you can use the search functionality through your MCP client:

1. Basic search: "Search for information about quantum computing"
2. Model-specific search: Use the web_search tool with a specific model parameter
3. Environment-based: Set DEFAULT_MODEL to change the default model globally

Deployment

This server can be deployed as a web service on platforms like Render.com, Heroku, or Railway.

Render.com Deployment

1. Fork or clone this repository to your GitHub account
2. Connect your GitHub repository to Render.com
3. Create a new Web Service with these settings:
   • Build Command: npm install && npm run build
   • Start Command: npm start
   • Environment Variables:
     • OPENROUTER_API_KEY: Your OpenRouter API key
     • NODE_ENV: production (optional)
     • DEFAULT_MODEL: Your preferred model (optional)

The server will automatically detect the deployment environment and run in web mode, providing both a documentation interface and REST API endpoints.

Web API Endpoints

When deployed as a web service:

• GET / - Documentation and status page
• GET /health - Health check endpoint
• POST /search - Web search endpoint

Example API usage:

curl -X POST https://your-deployment-url.com/search \
  -H "Content-Type: application/json" \
  -d '{"query": "latest AI developments", "model": "openai/gpt-4o:online"}'

Development

To modify or extend this server:

1. Edit src/index.ts
2. Run npm run build to compile
3. Restart your MCP client to reload the server

License

MIT License