A voice-enabled second brain inspired by the Moai of Rapa Nui — a stone guardian that watches over your own notes, connectors, and tools, and gives them a voice.
Your notes, connectors, and tools become a navigable 3D knowledge galaxy. Ask it questions and it answers from your own knowledge (or the web), flying the camera to the source; tell it to remember something and a new star is born live.
- Docs:
ARCHITECTURE.md(how it works)
py -3 build.py # generates/updates viewer/graph-data.js
py -3 server.py # starts at http://localhost:4700
py -3 server.py --dev # same, but the Dev console (local file operations) can be enabledDev mode is off-limits unless the server was started for it (--dev, or MOAI_DEV=1): otherwise nothing can turn it on over HTTP.
Open Chrome at http://localhost:4700. Paste your Anthropic API key into config.json (never in the chat).
Python 3 standard library only (build.py, server.py) + a single-page viewer (viewer/index.html) using 3d-force-graph and the browser's Web Speech API.
What leaves the machine: (1) your question plus relevant note excerpts → Anthropic API; (2) 3d-force-graph and Three.js downloaded from esm.sh on every load; (3) web search queries → Anthropic infrastructure; (4) voice recognition may use the browser's cloud service. Your notes themselves stay local.