June 29, 2007 — the first iPhone went on sale, and Apple redefined the phone.
19 years later, we hope to do the same for the keyboard.
Today, we’re launching our vision to rethink and redefine the way we interact with phones.
Excited to share what we’ve built. 🚀
The last major keyboard moment was in 2007,
when Apple put the keyboard on glass.
Nearly 20 years later, we're introducing:
Acti @openacti1, the Agentic Keyboard.
Not another AI keyboard that fixes grammar.
Not another voice keyboard that types faster.
An invisible agent in every text field.
It doesn't predict your next word.
It takes your next action. 👇
@oleg_kai@kimmonismus You're reading it right — a keyboard process can't hold an agent, so we didn't put it there. The keyboard's the trigger; the real work runs behind it, Siri-AI style: on-device + secure cloud. It's only 'autocomplete' if you cram the whole agent into the keyboard. We didn't.
@rakhul@AnatoliKopadze 3/ Because the heavy work lives server-side, we can also stay strict up front: only the minimal context the OS exposes, no side doors. Capability and privacy aren't a tradeoff here — the architecture gives us both
@rakhul@AnatoliKopadze 2/ And to be clear, we're not building a desktop agent. We're a keyboard first — powerful where it counts: the moment you type. Not a heavyweight, do-everything Codex or Claude. That scope is a deliberate choice, and it's the whole point
@rakhul@AnatoliKopadze 3/ And to be clear, we're not building a desktop agent. We're a keyboard first — powerful where it counts: the moment you type. Not a heavyweight, do-everything Codex or Claude. That scope is a deliberate choice, and it's the whole point.
@rakhul@AnatoliKopadze 2/ We run the same kind of architecture as iOS 27's Siri AI: on-device models and a secure cloud tier working together. The keyboard is the entry point — the real agent work runs behind it, with room to be powerful. We never bet on one sandboxed process to carry it all.