Orb — The Assistant That Talks First

Orb is a voice-first AI you actually own. You name it, you shape its personality, and you decide what brain runs behind it. The AI model is just an engine in the settings menu — the product is the assistant that gets things done in your life.

That’s the whole idea: the backend disappears. Once it’s set up, you don’t think about which model is answering. You talk to Orb, Orb helps, and it sounds the same whether it’s running fully offline on your phone, on a free API, or on a server you host yourself.

What makes it different

• Bring your own brain. On-device, free APIs, or your own machine — your choice, swappable any time. No single company is the point.
• Always works. Offline it runs on-device. If a capability isn’t available right now, Orb tells you plainly instead of failing silently.
• One personality, every backend. Switch the engine and the voice stays your Orb. Models are invisible infrastructure.
• Self-hostable. Run the optional Orb server on your own PC for the most power — and connect to it from anywhere.
• Private by default. On-device data stays on your device. A server you run is yours; you only ever connect to your own.

Start here

1. Getting Started — install the app and pick how it thinks.
2. Brains & Models — the tiers, and how Orb stays consistent.
3. Free APIs & Chaining — get effectively unlimited compute for free.
4. Self-Hosting — run your own Orb server.
5. Connect Your PC — link the app to your server, privately or from anywhere.
6. Local Inference — run a model on your gaming PC and stream it.
7. Connectors — calendar, reminders, and more.
8. FAQ — privacy, offline, cost, troubleshooting.

New to all this? Just install the app and talk to it. Everything below is optional power — Orb works out of the box on-device.