Zernote lets UX researchers re-run transcription on a user interview in the correct language while keeping existing research notes on the same record.
Run transcription again on the same interview. Pick the language for the new pass.
Your notes stay put. The transcript and anything built from it get rebuilt.
Start from the interview page, the API, or Claude via the interviews edit tools.
Wrong STT language yields garbage text. Speech in one language through the wrong ASR setting turns into noise. Summaries on top of that text can invent details.
Re-uploading the file was slow and risked losing notes you already wrote. Retranscribe keeps the interview record. Only the transcript side gets rebuilt.
A new pass uses the transcription pipeline again. Expect another token charge for that work.
If a summary was written on a bad transcript, re-check it after the new pass.
Topics