Open SourceLocal-first AI companion

The AI that
knows you.

L.U.N.A. runs on your machine. Voice, memory, meetings and desktop automation, on your desktop, in a browser, or on your phone.

Workflow

How L.U.N.A. works.

Four stages, entirely on your machine.

Speak or type

Chat through the UI or trigger Luna by voice. Accepts text, voice and attached files.

Context assembled

Memory facts, personality, calendar state, and activity history are injected into the prompt.

LLM processes

Your local Ollama model or any OpenAI-compatible API handles inference with a full context window.

Actions & response

Luna streams the reply, executes tools, opens widgets, updates memory, and delivers voice output.

See full workflow →

Built in the open.

Free, open-source, and always will be. Inspect every line, fork it, or contribute.