Your First Transcription
How to create your first transcription by recording or uploading audio.
Transcription Setup
To create your first transcription, click New Transcription in the transcribe pane on the homepage. You will be prompted to enter the Patient Name to begin.
Consultation Mode
You can choose to use consultation mode. This mode allows Letters to detect multiple speakers and name them accordingly in the transcription output.
Record and Upload Audio
Record a new transcription by clicking Start in the record menu of the audio pane.
Alternatively, upload an audio file from the upload menu.

Your Notes
You can also input your own notes to be used in the transcription for things you have not said out aloud.
Transcription Generated
Upon completion of a recording, a transcript will be generated in the transcript pane, which offers various editing and organisation options.
Detect and Name Speakers
Detect and Name Speakers
If using consultation mode, you can detect and name the speakers by clicking manage speakers.
Edit
Edit
Manually edit the generated transcription text.
Write Letter
Write Letter
Start a new Letter and automatically add the transcription as a source.
Action Menu
Action Menu
Intelligence
Intelligence allows you to customise how your transcription is interpreted and summarised. You can manage this from the Intelligence tab in the Transcription Output Pane.
You will be given the option to choose an Intelligence Type:
- Notes: A detailed clinical summary with structured headings (e.g. Presenting Complaint, History, Examination).
- Patient Summary: A simplified version of the notes, optimised for patient understanding.
AI Suggestions
Based on the content of your transcript, the Assistant will generate real-time suggestions—such as Disease Fact Sheets or Patient Action Plans—to support clinical workflows.
Custom Intelligence Types
You can also create your own Custom Type to match your preferred documentation style. Give it a name (e.g. SOAP Notes) and provide formatting instructions. Once saved, it becomes reusable across all future transcriptions.