Draft summaries for review
A structured draft from the transcript, which the clinician edits and approves.
Healthcare devices · In the consultation room
On-device transcription
Speech is turned into text on the device, with no audio sent to a server.
- Desktop
- Tablet
- Works offline
Desktop assistantOn this device · no internet needed
Transcribe the last ten minutes.
Done. 1,240 words, with speaker labels. Nothing was uploaded.
Label the speakers as clinician and patient.
Done. Each line is now labelled clinician or patient. Please check the labels.
Delete the recording but keep the text.
The audio is deleted from this device. The transcript is kept for your review.
The problem
Clinicians spend a large part of the day typing notes instead of looking at the patient. Cloud transcription sends conversations off the premises, which many organisations can’t allow. Rooms with poor connectivity make cloud tools unreliable anyway.
Who uses it
Doctors, nurses and therapists in consultation rooms, on the room PC or a tablet. Your clinical informatics team sets it up.
What you train it on
- Clinical vocabulary lists
- Drug and device names
- Your note templates
How it works on the device
- 1
Speech model on the device
A transcription model runs on the room PC or tablet, so audio is processed where it is spoken.
- 2
Your vocabulary
It is tuned on your terminology, drug names and abbreviations so the text needs fewer corrections.
- 3
The clinician stays in control
The transcript appears for the clinician to read, edit and keep or discard.
Why it runs on the device
Audio never leaves the room
No recording or transcript is sent to a server.
Works without a connection
Rooms with weak signal are no problem.
More time with the patient
Less typing during and after the consultation.
Questions
Only if you choose to keep it. By default the audio is discarded once transcribed, on the device.
In the languages you train it on. Accuracy is tested on your own recordings during the pilot.
It produces text. Moving it into your records system follows your existing process.
Related use cases
All use casesPut AI where your work happens.
Pick one team and one use case. We’ll have it running on your own devices in weeks.