Your AI tools forget you every conversation. Munkymind is an open-source personal context library: point it at your notes, and Claude, ChatGPT and other tools can draw on them whenever they need to.
Markdown, text, PDFs or GitHub repos. Munkymind reads them and builds a searchable library of who you are and what you're working on.
Connect Claude Desktop, claude.ai or ChatGPT through MCP. They pull the right pages when a question needs your context.
Every answer cites the notes it came from, and stale pages are flagged, so you can see why it said what it said.
Self-host with Docker. Your notes and library stay on your computer.
One OpenAI key is enough, or use Claude for answers. Want it fully local? Use Ollama and no key at all.
Apache 2.0 licensed. Read the code, change it, add your own connectors.
You need Docker, git and one OpenAI API key. Full checklist: Before you start.
git clone https://github.com/munkymind/munkymind.git
cd munkymind
cp .env.example .env # then put your OPENAI_API_KEY in .envmkdir -p notes && cp -r ~/path/to/your/notes/* notes/
docker compose up -ddocker compose exec api munkymind setup
Save the mm_sk_… key it prints. Then ask: docker compose exec api munkymind query --user yourname "What am I working on?"
Munkymind is at v0.2.0 and we're looking for early testers. Install it, try it with your real notes, and tell us how long it took and where you got stuck.