haiku.rag MCP
haiku-rag
Agentic RAG over your documents with citations (page, heading) and a built-in LanceDB. The MCP server starts with haiku-rag mcp --stdio
- 626 ★ on the repository
- Last commit
- MIT
Minimal access
Why this level
- An external embedding provider will receive the texts
- The full version pulls many dependencies
Install
- Works in
In your terminal, with SkillFoxx CLI
npx skillfoxx add mcp/haiku-ragDetects the agents on your machine, shows the access level and pins the version.
Other ways to install
Access and checks
How we reviewMinimal accessThe tool runs nothing and makes no network calls.
- An external embedding provider will receive the texts
- The full version pulls many dependencies
This is third-party code. Review the repository files before installing.
What it does
haiku.rag finds answers in your documents and gives citations with the page number and heading. The storage is built in, on LanceDB. Installation: `pip install haiku.rag` (full, with Docling, VoyageAI and Cohere embedders and a terminal interface) or `pip install haiku.rag-slim` (the core with extras you choose). MCP server: `haiku-rag mcp --stdio`. Version 0.92.0 from 02.10.2026, MIT license.
Who it is for. For people who want to ask their documents questions from an agent and get answers with page references.
Good fit when
- You need PDF and document search with citations
- You need a local RAG with no separate database
- You need an MCP server for document access
Not a fit when
- You cannot send texts to external embedding providers
- You need only a ready cloud service
Example request
Find the termination terms in the supplier contracts and give citations with page numbersMCP
- Transport
- stdio
- Authentication
- not required
Limitations
The document index is stored locally and access is minimal. If you choose an external embedding provider, the document text goes to it: check the settings before indexing. The full version pulls many dependencies and slim installs only the core.
How to disable. Remove the server entry from the client, the package with pip uninstall haiku.rag and the LanceDB database directory.
FAQ
What does haiku.rag do?
It finds answers in documents and gives citations with a page and heading.
How do I launch MCP?
haiku-rag mcp --stdio.
Where do the texts go?
The index is local but an external embedding provider will receive the text.
Discussion
No questions or reviews yet.
Related
MCP serversMCP Local RAG
mcp-local-rag
Local search over your documents, semantic and keyword, with no keys, Docker or external database. Works offline after the first model download
A CLI engine that compiles documents into a local structured wiki: a transparent alternative to plain RAG, reachable by an agent over MCP
On-premises RAG in containers with an MCP server: the agent searches your documents while data stays local
A CLI and MCP server that extracts structured knowledge from documents: graphs, hypergraphs, and fact search with sources
An MCP server for searching and downloading academic papers from arXiv, PubMed, bioRxiv, Semantic Scholar and other open sources
Official Pinecone MCP server: documentation search, index creation, record upserts and search from your agent
README in short
The README describes an agentic RAG on LanceDB with citations, installation through pip and an MCP server. MIT license.