Ambient Scribe
Record a consultation and let AI generate a structured clinical note automatically.
Open Scribe
Click Scribe in the sidebar. Select the patient from the dropdown at the top. You can also navigate directly with a patient pre-selected by clicking Try Scribe from their chart or the Sandbox Patient card.
To test without affecting real patient data, use the Sandbox Patient from the Patients page. It's designed for exactly this purpose.
Start recording
Click Start Recording. Your browser will request microphone access — allow it. The timer starts immediately. Conduct the consultation naturally — the AI listens to the full conversation.
There is no need to speak directly into the microphone. Normal room audio during a consultation is sufficient.
Stop and process
Click Stop when the consultation ends. The recording is compressed and sent for transcription. After a few seconds you'll see a structured consultation note generated from the conversation.
Processing time depends on the length of the recording. A typical 5-minute consultation takes 10–20 seconds to process.
Review and edit the note
The generated consultation note is shown in the editor. Review it for accuracy and make any corrections — the AI does not always catch everything perfectly. The note title is generated automatically but can be changed.
Save as preliminary, then finalize
Click Save to store the note as a preliminary draft. Preliminary notes are visible only to your care team — you (or a colleague) can keep editing and add to them across the visit. When the note is complete, click Finalize: this locks it against further edits and makes it visible to the patient in their records. The encounter date is recorded with the note.
Nothing reaches the patient until it's finalized. A nurse can capture a preliminary note (e.g. vitals and history at check-in) for a provider to review and finalize.
Upload an audio file instead
If you recorded the consultation externally (e.g. on your phone), click Upload Audio and select the file. The same transcription and note generation flow runs on the uploaded file.
Supported audio formats: MP3, M4A, WAV, WebM. Files are compressed before upload to reduce processing time.
Questions? contact@mercurie.ai