docs-mcp-rs

Name: docs-mcp-rs
Availability: InStock
Author: eslzzyl

Supports UI

by eslzzyl

High-performance documentation indexing and search MCP server with hybrid vector and full-text search.

0 stars

0 views

Works in:Cursor

Exposes:ToolsResources

What it does\ndocs-mcp-rs is a specialized MCP server that allows AI assistants to scrape, index, and search documentation websites. It uses a hybrid search approach—combining vector embeddings (via OpenAI or Google) and full-text search (FTS5)—to provide highly accurate retrieval of technical documentation.\n\n## Tools\n- `scrape_docs`: Scrapes and indexes a specified documentation URL.\n- `search_docs`: Performs hybrid search across indexed libraries for a specific query.\n- `list_libraries`: Lists all documentation sets currently indexed by the server.\n- `remove_library`: Deletes a specific indexed library from the local store.\n\n## Installation\nBuild from source and run in stdio mode:\n\n`bash\ngit clone https://github.com/Eslzzyl/docs-mcp-rs.git\ncd docs-mcp-rs\ncargo build --release\n./target/release/docs-mcp-rs serve --stdio\n`\n\n## Supported hosts\n- claude\n- cursor

cargo build --release && ./target/release/docs-mcp-rs serve --stdio

Choose your AI client and follow the steps below.

Cursor

Add the binary path to the MCP settings in Cursor.

Claude Desktop

Add the binary path to claude_desktop_config.json with the required environment variables (OPENAI_API_KEY).