On-device · nothing uploaded

Transcription for confidential audio
GDPR & HIPAA-friendly, no cloud

If you transcribe client calls, patient notes, legal interviews, or research with named participants, uploading that audio to a cloud service is often not allowed — by policy, contract, or law. Brethof Voice Pro runs the whole job on your own machine, so the compliant choice is also the simplest one.

Start the free 14-day trial

Not legal advice — but the audio genuinely never leaves your device.

The compliance problem

Cloud transcription uploads the one thing you can't share

A cloud transcription tool streams your recording to a third-party server to process it. For regulated work that's frequently a non-starter: GDPR data-minimisation and processor obligations, HIPAA for protected health information, attorney-client privilege, research ethics approvals, NDAs. The safest way to satisfy all of them is the same: don't send the audio anywhere. On-device transcription removes the third party from the equation entirely.

How on-device transcription helps

Private by design, not by setting

🔒

No cloud mode

There is no toggle that sends audio to a server for "better accuracy." Your CPU or GPU is the only processor. Nothing to misconfigure.

💾

Audio never hits disk

The buffer lives in RAM during transcription and is freed the moment text is produced — nothing to leak, nothing to recover.

📡

No telemetry

No usage stats, no crash phone-home. The only network calls are a licence check and the model downloads you trigger — documented and disableable.

Accuracy

Local doesn't mean lower quality

Transcription accuracy is competitive with the cloud's go-to model — ~4.5% word error rate on English (measured after a quick on-device fine-tune) versus Whisper Large-v3's 7.4% on the same audio — and you can fine-tune it further on your own voice and field vocabulary (medical, legal, technical) without anything leaving the machine. It transcribes 30 languages and can translate across 38.