Record with consent
The tool does not capture the room. You bring a file you already recorded.
Runs in this browser · Nothing uploaded · Weekly trial
Free interview transcription.
The tape stays in this tab.
Drop a consent-recorded interview. Speech to text runs here. No bot, no cloud account, no file on our servers.
Runs in this browser. The file never leaves this device.
Transcribe an interview without uploading
Research, hiring, and client interviews often cannot go to a cloud speech service. This page runs a public Whisper model in the tab. You remain responsible for consent, retention, and what you do with the text. The tool does not collect signatures or store the tape.
Private interview transcription
Download the model once. Then open the Network tab, drop the file, and confirm the audio is not in any request. After the cache is warm, the run can work with the network off. That check is the point of a local trial.
GDPR and Swiss nDSG for interviews
Swiss and EU research guidance treats free online speech-to-text as a processor risk. This page does not receive the file. It is still not a substitute for your own legal review, and it is not a certified transcript.
Qualitative interview transcription
Long tapes, two-person coding with speaker labels, and a dialect path belong on a Mac you control. The browser cap is 15 minutes and three runs a week.
Who is speaking?
Two-person interviews are easier to code with speaker labels. This browser pass does not diarize. If you need who-said-what, use Vemoir for Mac.
Swiss German transcription
Dialect interviews often come back as Hochdeutsch. That can be useful, but it is not a verified Swiss German model. Do not treat the wording as a certified transcript.
How to transcribe an interview
Drop the file here
MP3, M4A, WAV or MP4, under 15 minutes. Nothing is posted to Vemoir.
Confirm in the Network tab
After the model is cached, the audio should not appear in any request.
Review the text
Check names, numbers and anything you will quote. This is not a certified transcript.
Keep or delete the file yourself
Retention stays with you. We never received a copy.
Guides
Before you start
Is this a third-party processor under GDPR or nDSG?
The audio is not sent to Vemoir. The first model download comes from the published model host. After that, inference stays in the tab. This is not a substitute for your own legal review.
What about IRB or research policy?
Swiss research infrastructure has warned that free online speech-to-text is not safe for sensitive interviews. This page is the no-upload path. Long or dialect interviews belong on a Mac you control.
Do you keep a copy for training?
No. There is no upload, so there is nothing to train on. Telemetry, if you allow analytics cookies elsewhere on the site, is counts only.
Is the output a certified transcript?
No. It is a browser trial. Treat names, numbers and quotations as something you check.
Can I transcribe Swiss German interviews here?
You can drop the file. The public model usually writes Standard German. That is not a verified dialect engine.
Does this join the interview as a bot?
No. You record with consent on your side, then drop the file.
What file types work?
MP3, M4A, WAV and MP4, up to 15 minutes.
Where do longer interviews go?
Vemoir for Mac, with local models and speaker separation.
Keep the next interview on a Mac you control.
Vemoir for Mac transcribes locally, remembers speakers, and never joins the call as a bot.