Transcribes consultations and generates clinical documentation. Trains its own transcription models and small domain models from open datasets, using consented audio.