trackmcp
Back to directory

Agentic RAG for local and self-hosted document search: hybrid retrieval, reranking and multimodal RAG on embedded LanceDB, with Docling parsing and an MCP server

588 stars PythonOthers Updated Sep 4, 2026
mcp-serverraglancedbpydantic-aidoclingag-uiagentic-ragcode-executioncode-modecross-encoderdocument-aidocument-ingestiondocument-parsinghallucination-preventionmultimodal-ragoffline-llmrag-evaluationreciprocal-rank-fusionreranking

No README could be loaded. View the project on GitHub for full documentation.

Frequently asked questions

What is haiku.rag?

haiku.rag is Agentic RAG for local and self-hosted document search: hybrid retrieval, reranking and multimodal RAG on embedded LanceDB, with Docling parsing and an MCP server

How do I install haiku.rag?

Open the GitHub repository and follow its README. Most MCP servers are added to your client's MCP config, then called by your agent.

Is haiku.rag open source?

Yes — it is hosted on GitHub at https://github.com/ggozad/haiku.rag and has 588 stars.

Related MCP tools

Run your own MCP server? See who uses it and what to fix.

Measure it with TrackMCP