lyonzin/knowledge-rag

Overview

knowledge-rag indexes documentation for instant retrieval from Claude Code. It provides 12 MCP tools, supports 20 document formats, and uses hybrid search plus reranking. The system runs entirely locally with zero servers and no API keys.

Best for

Best for
Developers using Claude Code who need private, local document retrieval

Use cases

• Query project documentation directly from Claude Code
• Enable Claude Code to access domain-specific knowledge without external services
• Integrate local technical guides into coding sessions

How to use

Install

npx -y knowledge-rag                    # NPM — zero setup, auto-manages Python venv

Tools exposed

• max_results
• hybrid_alpha
• min_score
• snippet_mode
• full_rebuild
• delete_file
• test_cases

Tested with

Claude Desktop, Claude Code, Cursor, Windsurf, VS Code

Notes

knowledge-rag indexes documentation for instant retrieval from Claude Code. It provides 12 MCP tools, supports 20 document formats, and uses hybrid search plus reranking. The system runs entirely locally with zero servers and no API keys.

88 stars on GitHub. Last updated 2026-05-26. Licensed MIT.

Use cases

• Query project documentation directly from Claude Code
• Enable Claude Code to access domain-specific knowledge without external services
• Integrate local technical guides into coding sessions

Pros

• Fully local, no server setup or API keys needed
• Supports a wide range of document formats (20 parsers)
• Combines hybrid search with reranking for better retrieval

Cons

• Designed specifically for Claude Code, limiting use with other tools
• Requires Python environment and manual document dropping
• Relatively small community (88 stars) may mean limited ongoing support

Indexed from awesome-mcp-servers-punkpeye and enriched against its public facts.