Browser-only speech capture to help you grab the important parts of a session. Nothing is uploaded; your audio and transcripts stay in this tab.
Press Start Recording and speak normally — people, roles, songs, and anything you add below are picked out of the transcript. Add names / song titles / nicknames under Artists, Songs, or Other, then press Apply: they become trigger words and are also boosted inside the speech model itself. If you're in a noisy room, open Audio settings and use the input toggles.
One per line, as Label: alias, alias — e.g.
Maya: maya, may or
Song 3: song three, track 3.