pi-mnemosyne

Pi extension and skill for Mnemosyne — local-first, SQLite-backed AI memory.

@mnemosyne-oss/pi-mnemosyne brings Mnemosyne into the Pi coding agent through a native Pi extension and skill.

Features

• Native Pi tools for remembering, recalling, and forgetting memories.
• Local-first — all data stays in a SQLite database on your machine.
• Zero-config once installed.
• Pi skill included for progressive disclosure and best practices.

Install

1. Install the Mnemosyne CLI:

pip install mnemosyne-memory

2. Install the Pi package:

pi install npm:@mnemosyne-oss/pi-mnemosyne

Or add it to .pi/settings.json for a project:

{
  "packages": [
    "npm:@mnemosyne-oss/pi-mnemosyne"
  ]
}

3. Restart Pi (/reload or restart the session).

Tools

Tool	Purpose
mnemosyne_remember	Store a memory
mnemosyne_recall	Search memories by semantic similarity
mnemosyne_forget	Delete a memory by ID
mnemosyne_stats	Show memory statistics
mnemosyne_sleep	Consolidate old memories into summaries

Quickstart

After installing, try:

/skill:mnemosyne

Then ask:

Remember that I prefer vitest over jest for testing.
Recall my testing preferences.

Development

git clone https://github.com/mnemosyne-oss/pi-mnemosyne.git
cd pi-mnemosyne
npm install
npm run typecheck
npm test

Releasing

1. Bump version in package.json.
2. Commit and push.
3. Tag and push:

   git tag -a v0.1.0 -m "Release v0.1.0"
   git push origin v0.1.0
   

4. The release workflow publishes to npm.

Make sure the NPM_TOKEN secret is set in the GitHub repository settings.

License

MIT © Mnemosyne OSS