Developers
API
The same agent that talks on this site, from your code. The API is not public yet — the site runs on it, and we are cleaning it up before opening the door.
What it will do
Talk
Send a message, get the answer with the emotion and, if you want, the voice file. One call, one energy.
Memory
Read and write what Lumi remembers about a person — the same memory every face shares.
Pictures
Ask for a picture in the person’s style and get it into their library.
Device
See which Lumi One is online, what it shows, and send it a scene.
How it will look
POST https://api.lumi-numi.com/v1/talk
Authorization: Bearer <your key>
{ "face": "cat", "text": "What should I cook tonight?" }
→ { "answer": "…", "emotion": "curious", "voice": "https://…/voice/….mp3", "energy": { "spent": 2 } }Keys will be tied to your account and spend its energy. Same prices as talking on the site.
Want in early?
Write to us with a line about what you are building. We hand out keys by hand for now.
[email protected]Using an agent framework? See the MCP server.