Clinicians spend a large part of their day writing notes. This platform lets them talk instead: upload a recording, dictate live, or let ambient AI listen to the consultation, and receive a structured clinical document ready for review. It follows our earlier work with XStek on their medical transcription platform.
The challenge
Medical notes have to be accurate, follow a clinical structure and protect patient privacy. Generic transcription tools return a wall of text that still needs heavy editing.
What we built
We combined speech recognition with large language models that organise the transcript into the sections clinicians expect, and wrapped it in a simple web workflow with three ways to capture audio: file upload, dictation and ambient listening.
The result
Clinicians get a structured first draft of their notes in minutes instead of typing them from scratch, and can review and correct it in one place.
Key features
- File transcription for recorded consultations
- Ambient AI mode that listens during the visit
- Live dictation mode
- Speech turned into structured clinical notes
- Review and edit before notes are finalised
Outcomes
- Less time spent typing notes
- Consistent note structure across clinicians
- Three capture modes to suit how each clinician works
