McpDocServer

# McpDocServer A MCP protocol-based development documentation server designed for various development framework documentation. It provides multi-threaded document crawling, local document loading, keyword search, and document detail retrieval functionality. ## Core Features Demonstration ### 1. Document Crawling Demo  *Complete document crawling process from configuration to executing `npm run crawl`* ### 2. MCP Server Call Demo  *Process of querying API and obtaining accurate documentation results in Cursor* ## Solving Cursor's Hallucination Issues When using Cursor for framework development, AI often encounters "hallucination" problems due to imprecise understanding of framework APIs: - **Accuracy Issues**: AI might recommend non-existent or outdated framework APIs and components - **Version Confusion**: Mixing APIs from different versions, leading to non-functional code - **Parameter Errors**: Inaccurate understanding of method parameters, especially for framework-specific features - **Compatibility Misjudgment**: Unable to accurately determine API compatibility across different environments or platforms This MCP server effectively solves these issues by providing precise documentation retrieval: - **Real-time Precise Queries**: Direct access to the latest accurate API information from official documentation - **Context Association**: Related API and component documentation displayed together for complete reference - **Precise Parameter Matching**: Complete method signatures and parameter lists to eliminate parameter errors - **Cross-platform Compatibility Markers**: Clear identification of API compatibility across platforms - **Example Code**: Official example code provided to ensure correct usage By integrating this MCP server, you can significantly improve Cursor's accuracy and efficiency in framework development, avoiding development obstacles caused by "hallucinations". ## Features - Support for loading framework documentation data from local JSON files - Powerful documentation search functionality - Document detail query - Automatic recognition of available documentation sources - Support for targeted queries of specific documentation sources - Support for crawling external documentation, automatically converting to locally usable format - Support for documentation reloading (triggered by searching "reload") ## Directory Structure ``` / ├── server.js # Server entry file ├── docs/ # Documentation data directory │ ├── taro-docs.json # Taro framework documentation │ └── taroify-docs.json # Taroify component library documentation ├── scripts/ # Scripts directory │ └── crawl.js # Documentation crawling script ├── tests/ # Test directory │ └── mcp.test.js # MCP test script ├── config/ # Configuration files directory │ └── doc-sources.js # Documentation source configuration └── package.json # Project configuration ``` ### Installation and Running If you have Chrome browser installed locally and want puppeteer to use your existing version, you can set the PUPPETEER_SKIP_DOWNLOAD environment variable: macOS/Linux: ```bash export PUPPETEER_SKIP_DOWNLOAD=true npm install ``` Windows (Command Prompt): ```bash set PUPPETEER_SKIP_DOWNLOAD=true npm install ``` Windows (PowerShell): ```bash $env:PUPPETEER_SKIP_DOWNLOAD = $true npm install ``` ### Creating Crawler Configuration Create a `doc-sources.js` file in the `config` directory following this format: ```javascript // config/doc-sources.js // Documentation source configuration export const docSources = [ { // Documentation source name - used as source parameter when searching name: "taro", // Documentation website base URL url: "https://docs.taro.zone/docs", // Include patterns - specify URL paths to crawl (empty array means all pages) includePatterns: [ ], // Exclude patterns - specify URL paths not to crawl (supports regular expressions) excludePatterns: [ /\d\.x/, // Exclude version number pages /apis/ // Exclude API pages ] } // ... more sources ]; // Crawler global configuration export const crawlerConfig = { maxConcurrency: 40, pageLoadTimeout: 30000, contentLoadTimeout: 5000, headless: false, maxRetries: 3, retryDelay: 2000, requestDelay: 1000 }; ``` ### Running the Crawler After configuration, execute the following command to start the crawler: ```bash npm run crawl ``` ### Starting the MCP Server ```bash npm start ```