Private interview transcription
Create a draft you can check against the recording before quoting or sharing it.
Verify quotations against the original audio. Local processing keeps the selected recording on this device; speaker names and attribution still need manual review.
Prepare a draft for review
Keep the original
Preserve an unchanged recording. Keep it available for playback while you review the transcript.
Choose the file and language
Select the saved interview and its spoken language. The default Tiny model uses WebAssembly; model and backend options are available in advanced settings.
Transcribe, then inspect
Let the recording finish. Edit timestamped segments in the workspace and return to the audio whenever wording or attribution is uncertain.
Before a quotation leaves the draft
Check the exact words
Verify names, numbers, technical terms and negations. A missing word can reverse the meaning of a sentence.
Confirm the speaker
Whisper Web does not identify speakers. Add labels manually and replay overlapping speech before attributing a quotation.
Read the surrounding exchange
Review the question and surrounding answer before selecting an excerpt. Share only the text needed by the next reviewer.
Manage the recording and its copies
Confirm recording permission and the sharing rules for your interview. Local transcription avoids a media upload to Whisper Web; device access, backups and exported copies still need your attention. The application and model download over the network.
Use standard transcription up to 300 MB and 20 minutes, or the large-file tool for a local file up to 1 GB and 1 hour. Keep the tab open during processing and export the corrected text before leaving.
Common questions
Is local processing enough for a confidential interview?
It avoids uploading the selected recording to Whisper Web. Confidentiality also depends on who can access the device, where copies are backed up and how the exported transcript is shared.
Can I publish the automatic transcript as a quotation?
Check the words and attribution against the original recording first. Automatic text is a draft, including passages that look fluent.
Can I replay an interview from saved history?
Choose the original recording again to enable playback. Transcription history saves text and timestamps, not the source media.