knowledge-base-mcp

# Knowledge Base MCP Service A knowledge base retrieval system based on the MCP (Model Context Protocol) protocol, providing integration capabilities with the Dify knowledge base. ## Project Introduction The Knowledge Base MCP Service is a lightweight service that connects your Dify knowledge base to MCP-supported AI assistants (such as Claude AI) via the MCP protocol. This service supports various transport methods, including STDIO, HTTP, and SSE, making it suitable for different application scenarios. ### Key Features - Supports the MCP protocol for seamless integration with AI assistants like Claude - Provides tools for parsing knowledge base IDs and retrieving knowledge base documents - Supports multiple transport methods: STDIO, HTTP, and SSE - Caching mechanism improves query efficiency - Supports simultaneous queries across multiple knowledge bases - Simple and easy-to-use API interface ## Technical Architecture  The system consists of the following main components: 1. MCP Server: Handles communication with AI assistants 2. Dify Service: Interacts with the Dify API to retrieve knowledge base content 3. Tool Service: Provides MCP tool implementations, including parsing knowledge base IDs and retrieving knowledge base documents ## Installation and Setup ### Environment Requirements - Node.js 18+ - TypeScript - A valid Dify API key ### Installation Steps 1. Clone the repository ```bash git clone https://github.com/yourusername/knowledge-base-mcp.git cd knowledge-base-mcp ``` 2. Install dependencies ```bash pnpm install # or use npm npm install ``` 3. Configure environment variables Copy `.env.example` to `.env` and fill in the necessary configuration information: ```bash cp .env.example .env ``` Edit the `.env` file to set the Dify API key and other configurations. 4. Configure Docker ```bash cp Dockerfile.example Dockerfile ``` Edit the `Dockerfile` to set the Dify API key and other configurations. ### Configuration Description The following parameters can be configured in the `.env` file: | Variable Name | Description | Default Value | |---------------|-------------|---------------| | DIFY_BASE_URL | Dify API Base URL | https://api.dify.ai/v1 | | DIFY_API_KEY | Dify API Key | - | | TRANSPORT_TYPE | Transport type (stdio/http/sse) | stdio | | PORT | HTTP/SSE service port | 3000 | | CACHE_TTL | Cache time-to-live (milliseconds) | 3600000 (1 hour) | | CACHE_CLEANUP | Cache cleanup interval (milliseconds) | 300000 (5 minutes) | | ALLOWED_ORIGINS | Allowed CORS origins (comma-separated) | * | ## Build and Run ### Build the Project ```bash pnpm run build ``` ### Run the Service You can run the service in the following ways: #### STDIO Mode (default) ```bash pnpm start # or pnpm run dev ``` #### HTTP Mode ```bash pnpm run dev:http # or node dist/index.js --transport http --port 8008 ``` #### SSE Mode ```bash pnpm run dev:sse # or node dist/index.js --transport sse --port 8006 ``` ### Command Line Parameters The service supports the following command line parameters: - `--transport <stdio|http|sse>`: Specify the transport type - `--port <number>`: Specify the HTTP/SSE service port - `--dify-api-key <string>`: Specify the Dify API key (overrides environment variable settings) ## License MIT License