POST PRODUCTION
Give every voice a name.
Match speakers to your existing transcript. Listen, review and export — with your wording and timecodes preserved.
Private workspace
Sign in to access your recordings and transcripts.
Model access
Once only: accept Community-1 access, then create a Hugging Face Read token. Your token stays on this server and is used to download the model. Your recordings are processed here.
01 / ADD EPISODES
Upload one or several episodes
02 / YOUR WORKSPACE
Episodes
03 / LISTEN AND CONFIRM
Analysis runs in the background. You can leave this page and return later.
Choose a sample or a transcript cue to listen.
Name the voices
Listen to several samples for each voice. The app assigns clear cues to the voice with sustained speech; brief reactions are kept as evidence but excluded from the exported name. Review the uncertain cues below.
Changing the offset recalculates automatic assignments. Saved manual choices are preserved when you improve suggestions.
Take your transcript back to the edit
Clear speaker assignments are accepted automatically. Drafts mark uncertain cues. Final export needs names, confirmed timing and any uncertain cues resolved.