On-device transcription means the speech-to-text model runs on your own computer instead of a remote server. For meeting notes — some of the most sensitive data a company handles — that distinction is everything.
Cloud vs on-device
- Cloud: your audio is uploaded, transcribed remotely, and stored by the vendor
- On-device: your audio is transcribed locally and never leaves the machine
Why it changes your security review
If no data leaves the device, there's no third-party data-processing agreement to negotiate, no sub-processor list to vet, and no breach surface at the vendor. Tools that would otherwise be blocked sail through review.
Is it accurate enough?
Yes. Modern browsers can run OpenAI's Whisper via WebGPU at near-real-time speeds, and Whisper is the same family of model the cloud tools use. You get comparable accuracy without the upload.