speakers
Who speaks in a recording and when: voices told apart and matched against those this workspace has named, stored on the item. name stores a voice print that recognises that person across the workspace’s recordings; the user’s Terms require that person’s permission where the law does. unname takes a name off this recording; merge joins one person read as two speakers. too_short marks too little speech to match, and sounds_like a probable match to confirm by naming. Times are seconds into the file. transcript: true labels stored sentences by speaker. For the words, transcribe_media. First read: cloud processing.
POST
speakers
Authorizations
Headers
Body
application/json

