Setting up
What arrives, how it is set up, and how skills get onto it.
1. It boots and starts listening
Jarvis runs as a background service. Plug in power, a microphone and a speaker, and it is already listening. There is no app to open.
2. Open the dashboard
On the same network, open the box's address in a browser. Set your WiFi, pick your microphone and speaker, and add your API key.
3. Install the skills you want
Download from the library, upload on the dashboard's Library page. Some skills want a small ESP32 board; most just work.
4. Teach it your voice and face
Enrol yourself so it can tell who is speaking, greet people it knows, and hold reminders until you are actually in the room.
How a skill works
A skill is a folder. It holds a manifest describing what it can do and the phrases that should route to it, a handler that does the work, and an optional composer that turns a result into something worth saying out loud. That's the whole contract.
my_skill/
skill.json what it can do, and what to route to it
handler.py execute(action, parameters, config, manifest)
composer.py optional: turn a result into a spoken reply
store.json optional: hardware, wiring and setup notesZip that folder and upload it on your dashboard's Library page. Your own skills sit alongside the ones from this library, and nothing here can see or touch them.
Your data does not come to us
There is no account and no server holding your conversations. Recordings, transcripts, contacts and API keys stay on your device. This website serves files; it never talks to your Jarvis, and your Jarvis never talks to it unless you download something.
Firmware
Sketches for the boards some skills use.