I built an MCP on top of OOREP
Hi there đź‘‹
First of all, thank you for releasing and maintaining OOREP. It’s an amazing resource for homeopathic repertorisation and materia medica work. I wanted to let you know that I’ve built a community integration for OOREP for use with AI assistants and agent frameworks:
-
Project:
Dhi13man/oorep-mcp -
npm package:
oorep-mcp - Protocol: Model Context Protocol (MCP) - used by tools like Claude Desktop, Claude Code, various CLIs, and agent frameworks
In short, oorep-mcp is:
A Model Context Protocol server and TypeScript client SDK that lets AI assistants query OOREP (the public instance at https://www.oorep.com) for:
- Repertory searches (12+ repertories)
- Materia medica searches
- Remedy information lookup
- Listing available repertories and materia medicas
- Guided workflows for symptom analysis, remedy comparison, and case repertorisation
Some quick implementation notes:
- By default it talks to the public OOREP instance (
https://www.oorep.com) via HTTP. - It does not embed or redistribute the OOREP codebase or database; it’s a thin integration layer that forwards structured queries and returns structured results.
- It’s MIT-licensed and clearly positioned as an independent, community-maintained project, not an official OOREP component.
Why I’m opening this issue
I thought this integration might be useful to other users of OOREP who are experimenting with AI assistants (Claude, OpenAI via function calling, Vercel AI SDK, LangChain/LangGraph, etc.) and would benefit from a ready-made way to access OOREP programmatically via those tools.
I wanted to ask if you would be open to adding a small note or “community integrations” footnote in the OOREP repository’s README.md that links to this project.
For example, something along the lines of (wording completely up to you, of course):
## Community integrations
- [oorep-mcp](https://github.com/Dhi13man/oorep-mcp) – Community-maintained Model Context Protocol (MCP) server and TypeScript SDK that lets AI assistants (e.g. Claude, agent frameworks) query the public OOREP instance programmatically.
> Note: This project is not officially affiliated with or endorsed by the OOREP project; it is a third-party integration.
If you are open to this, I can raise the PR myself as well.