Voice Lab

Turn speech into text, match words, change colors, or see the frequencies in your voice.

Microphone off

Your words will appear here.

Microphone off.

On-device recognition only

Downloads an English model (about 40 MB) on first use. Audio stays on your device; no browser speech service or account is needed.

Try with text instead

Example transcript

Behind the experiment

Word matching adapts Cleo’s Club’s transcript comparison. Sound waves adapts Snoopi’s microphone visualizer. On-device speech uses Vosk and an English model downloaded on first use. The optional browser speech service may process audio online. Sound waves uses Web Audio locally.

Listening stops when you leave the page, hide the tab, or press Stop. Speech sessions also stop after one minute. Nothing is saved by this site.

Vosk Browser (Apache-2.0) ↗English model ↗

Cleo’s Club ↗Snoopi ↗

Face Recognition Fields