🧪 Distill¶

Privacy-first team memory for Claude Code.

An MCP server that gives Claude Code a shared team knowledge base. A local LLM transforms your raw input into anonymous, factual knowledge before anything leaves your device.

No author. No frustration. No names. Just a clean, reusable fact.

Distill demo — raw thought to clean team fact

flowchart LR
    A[You type freely] --> B[Ollama on your Mac]
    B --> C{Review}
    C -- approve --> D[Team DB]
    C -- reject --> E[Discarded]

Quick start¶

pip install distill-mcp
ollama pull gemma3:4b && ollama pull nomic-embed-text
claude mcp add distill -- python -m distill_mcp

Then in Claude Code:

You:    "Remember that we chose gRPC over REST for streaming and type safety."
Claude: [distills locally, shows preview, waits for approval, saves]

Documentation¶

This site follows the Diataxis framework:

Tutorials — Step-by-step guides for getting started
How-To Guides — Task-oriented recipes for specific goals
Reference — Technical details: tools, configuration, ports
Explanation — Architecture, privacy model, design decisions