Browser demo · no database
STT-1
Speech to text, first pass. Two text boxes, one popup — same write-back idea as the date picker. Access cannot do this. The browser can (Chrome or Edge, microphone permission).
Two boxes
Not linked to a database. Each Speak opens the same popup and can write back to
that box (same idea as date-picker Go!).
Only one tab can use the microphone at a time — close any other tab that shows a red recording dot first.
What this demo shows
A speech-to-text popup that returns text to the field that opened it — STT-1. Not an Access file. Access has no speech recognition.
- Calling control — Text 1 and Text 2 share one popup. Use this writes back to the box that opened it (same pass-back as the date picker).
- Microphone — the browser’s Speech Recognition API (Chrome or Edge). Press Mic, speak, words appear in the large area. Press Stop when you are done. Only one tab can record at a time: if another tab shows a red recording dot, close it first or this page will not hear you.
- Close without sending — Close leaves the host box unchanged. You choose whether the transcript goes back.
- No database — two unbound boxes. This page only tests whether a speech interface is possible.
- Ask Grok — after a transcript, Grok fills first name, last name, date of birth, phone, and email. The API key stays on the server. You check the card, then Use this.
- Password — the whole STT-1 demo (Speak and Ask Grok) is locked. Hiding the popup alone would not stop a direct call to the extractor. The password is checked on the server.
Will this work when the data is in a database?
Yes. The host field would be ordinary text (or a memo). The popup is UI. The same sentence would save like any other typed value. Firefox and Safari often have no Speech Recognition API; Chrome and Edge do.
Needs a microphone permission prompt. Works on https or localhost, not usually on file://.