Model Context Protocol (MCP) Server
by infinyte
The Model Context Protocol (MCP) Server provides a unified API for multiple AI model providers. It simplifies integration and management of various AI models within a single application.
Last updated: N/A
What is Model Context Protocol (MCP) Server?
The MCP Server is a server implementation for the Model Context Protocol. It acts as a unified API layer, allowing developers to interact with multiple AI model providers (e.g., Anthropic, OpenAI) through a single, consistent interface.
How to use Model Context Protocol (MCP) Server?
To use the MCP Server, first clone the repository and install the dependencies using npm install. Then, run the interactive setup script (npm run setup) to configure API keys. Configure MongoDB using one of the provided options. Finally, start the server using npm start or one of the other startup options. Interact with the server through the API endpoints, such as /mcp/:provider for making requests to AI models and /tools/available for discovering available tools.
Key features of Model Context Protocol (MCP) Server
Unified API for multiple AI providers
Support for chat completions and legacy completions
Tool calling support
Context/system message handling
Environment-based configuration
MongoDB database for persistence and state management
Tool execution history and analytics
Use cases of Model Context Protocol (MCP) Server
Building AI-powered applications that leverage multiple AI models
Creating a consistent interface for interacting with different AI providers
Implementing tool calling functionality for AI models
Managing context and system messages for AI models
FAQ from Model Context Protocol (MCP) Server
How do I add a new AI provider?
How do I add a new AI provider?
Add the provider's SDK, create a handler function in server.js, and add a case in the main route handler.
How do I add a new tool?
How do I add a new tool?
Create a new tool implementation in /src/tools, add the tool definition to tool-definitions.js, update the tool execution functions in server.js, and add new API endpoints if needed.
What API keys do I need to configure?
What API keys do I need to configure?
You need API keys for the AI providers you want to use (e.g., ANTHROPIC_API_KEY, OPENAI_API_KEY). You may also need keys for web search (GOOGLE_CSE_API_KEY, GOOGLE_CSE_ID, BING_SEARCH_API_KEY) and image generation (STABILITY_API_KEY).
How do I run the server in production?
How do I run the server in production?
Use the PM2 process manager with the npm run pm2:start:prod command. This ensures the server restarts automatically if it crashes and is optimized for production environments.
How do I discover available tools?
How do I discover available tools?
Use the GET /tools/available endpoint. It supports multiple formats (JSON, YAML, HTML) and provides filtering by category, provider, and search terms.