WebAI StudioSpeak to Fill
SpeechSay "table for four next Friday at seven, outside" and watch the booking form fill itself — speech in, structured data out.
Why On-Device?
Voice hits the on-device recognizer, the transcript hits the on-device LLM with a JSON schema, and the form fills — no audio or personal details ever reach a server.
Interactive Demo
Tap and say something like
"Table for four next Friday at seven, outside."
Utterances refine the form — try the corrections after the first one.
The confirm button unlocks once name, date, and party size are filled.
Implementation Code
Keep exploring
56 demos, all running on-device.
Tongue Twister Trial
Say "she sells seashells" as fast as you can — the on-device recognizer scores how much of it survived.
Jargon Dictation
Dictate product names the recognizer has never heard — then boost them with contextual biasing and watch the diff.
Dictate & Polish
Speak your messy first draft, then let the Proofreader and Rewriter turn it into publishable text.