Overview
Create a new session to capture a consultation by recording or uploading audio, transcribing it, and generating an AI summary you can edit, download, or push to your EHR.
Prerequisites
Use a modern browser, ideally Google Chrome.
Allow microphone permission when prompted.
Optional: Set up your default AI model and default microphone in Settings → General.
Step‑by‑step
Open New Session
Go to New Session. Give the session a helpful name if desired.Pick a template
Select a favourite or default template (e.g., SOAP). You can change later and regenerate.Select or enter a patient
Link an EHR patient (if integrated) or enter details manually. You can link later before pushing.Attach context files (optional)
Add supporting files (PDF, DOCX, TXT, MD, JPG, JPEG, PNG; ≤3 MB). You can select up to 5 files to guide the AI.Choose an input method
Transcribe: Record the live consultation. Use Pause/Resume as needed.
Upload: Upload a pre‑recorded file to transcribe.
Dictate: Speak freely to produce a note quickly.
Generate the summary
Once a transcript is ready, click Generate Summary. Edit as needed. Versions are kept.Copy or download
Use Copy (with or without preserved line breaks) or Download (DOCX with your letterhead & signature).Complete the session
Click Complete to save and start fresh. If a recording is paused, you’ll be asked to confirm completing without saving the current recording.
Tips & best practice
Watch the Logs panel for a transparent timeline of uploads, transcription, and generation.
Long recordings are supported; for very long consultations consider splitting into parts.
Context files help tailor output—keep them concise and relevant.
Troubleshooting
No transcript created: Ensure mic permission is allowed, the correct device is selected, and speech is present. Silent audio may be skipped.
Upload failed: Keep files ≤3 MB and use supported formats. Check connection; uploads retry when possible.
Navigation warning: If transcription or generation is running, you will see a “processing in progress” warning when leaving the page.
