Medical transcriptionMedical transcription that knows the vocabulary

A general transcription model hears “metoprolol” and writes what it thinks it heard. Medical transcription lives or dies on exactly those words: the drug names, dosages, anatomical terms, and shorthand that a clinician says without slowing down. Hushscript’s medical mode switches that file’s recognition to a model built for clinical language, and a saved dictionary of your own terms narrows it further – while the recording itself is deleted the moment the transcript is ready.

Dictation and interviews, one workflow

Medical transcription is really two jobs. A clinician dictating notes alone can record straight into the app’s built-in recorder and submit without ever handling a file. A recorded clinical interview or case discussion uploads like any audio, and speaker labels separate the voices – interviewer from participant – without naming anyone. Either way, medical mode is a per-file switch: turn it on for the dictation, leave it off for the staff meeting, and only the clinical file pays the clinical rate.

The vocabulary, handled twice

Medical mode changes which model listens; a saved dictionary changes what it listens for. Build the dictionary once from your practice’s or study’s actual terminology – bulk-imported from a CSV or TSV, organized into categories, spelling variants attached to each canonical term – and mark it default so it rides along on every upload. The combination is what gets a medication history transcribed right on the first pass, with find-and-replace catching the stragglers everywhere at once.

The privacy trade-off, stated plainly

For patient-adjacent recordings there are two configurations, and they don’t combine. A normal job with medical mode gives you the clinical model and an encrypted transcript under your retention rules (auto-delete after 7 to 365 days, or per-transcript overrides), with dedicated deletion flows returning a receipt. Private mode gives you the shortest possible trail – no transcript saved to the account, just a password-protected archive – but runs speech recognition only, skipping medical mode and translation. Pick per recording. In both cases the audio itself is deleted the moment processing ends, is never used for training, and EU, EEA, Swiss, and UK audio is processed in the EU.

What medical transcription costs

Medical mode doubles the billed minutes for its file: a 20-minute dictation bills as 40 minutes, which on the largest pack ($49.99 for 100 hours, about $0.50 an hour) comes to roughly $0.33. A week of daily 20-minute dictations lands around $1.65. There’s no subscription and no monthly minimum – the quiet week costs nothing, and unused minutes stay valid for a year.

Everything on this page describes capabilities, not compliance status. Hushscript doesn’t claim regulatory standing; assess these capabilities against your organization’s own requirements before regulated use.