Chatterbox Labs
About
We got tired of AI that only lives in glass rectangles — and of “cloud-required” gadgets that die offline. So we built a box with its own NPU.
Origin
Chatterbox Labs started from a simple irritation: every assistant demo needed a phone, a tab, and a live connection. Desk work doesn’t. We wanted ambient conversation with honest defaults — local first, cloud when asked, mute that means mute.
Slogan stuck early: chatter your box. Slightly ridiculous. Still right.
What we build
- ChatterBox hardware (ChatterCore 12 TOPS NPU)
- Chatter-Local 1.8B — the free unlimited on-device model
- ChatterCloud Boost — optional metered escalation for hard turns
- Personality packs as first-class product, not a prompt dropdown
What we believe
- Voice should not require a screen.
- Basic AI should work offline.
- Cloud pricing must say what a “turn” is.
- Hardware mute is non-negotiable.
- If it only works in a demo video, it doesn’t ship.
The team
Small design-engineering shop: tactile hardware, edge ML, and writers who refuse generic assistant copy. Shipping in public toward founding batch.
Contact
Press / partnerships: hello@chatterbox.example
Support (at ship): companion app