Doxy.me and AI Scribes: Best Practices

Doxy.me is a HIPAA-compliant video platform built for behavioral health clinicians. Scriber has a real Doxy.me integration that lets you capture audio directly from your Doxy.me session and generate notes without an extra recording step.

This is the difference between "AI scribe and Doxy.me" and "AI scribe integrated with Doxy.me."

What Scriber's Doxy.me integration does

The standard workflow (without integration):

Clinician opens Doxy.me and starts a telehealth session.
Clinician starts a separate recording device or app.
Visit happens across two systems (Doxy.me for video, separate recorder for audio).
Visit ends.
Clinician manually uploads the recording to the AI scribe.
AI scribe generates note.

This works, but it requires manual steps and creates a separate audio file you have to manage.

The Scriber + Doxy.me integrated workflow:

Clinician opens Doxy.me.
Clinician starts Scriber inside Doxy.me (via the Scriber widget or native Doxy.me integration).
Patient is informed the session is being recorded for documentation purposes.
Visit happens with Scriber listening to the Doxy.me audio.
Visit ends.
Scriber generates the note automatically.
Clinician reviews, edits, and signs the note.
Audio is discarded.

No manual upload, no separate recording device, no audio file to manage. The Doxy.me session is the source, Scriber captures it, and the note appears in your Scriber workspace.

How audio flows from Doxy.me to Scriber

This is a technical detail that comes up in security reviews: how does Scriber capture audio from a Doxy.me session?

The answer: via getDisplayMedia and browser audio capture.

Doxy.me's WebRTC video stream does not expose the audio tracks directly to third-party integrations — that is a security feature. So Scriber uses the same approach every other AI scribe and recording app uses: it requests permission to capture the browser's audio output and the microphone input via the standard getDisplayMedia API.

When you start Scriber in Doxy.me:

Your browser asks: "Scriber is requesting permission to record audio from this session."
You grant permission (or deny it).
Scriber captures the audio flowing through your browser during the call.
The audio is sent to Scriber's encrypted processing pipeline.

This is the same mechanism used by Zoom recording, OBS Studio, and every other browser-based recording tool. It is not special or hidden — it is browser-native, secure, and standard.

Why not just ask Doxy.me for the audio directly? Doxy.me's architecture does not expose WebRTC tracks to extensions or integrations. The platform is designed to prevent unauthorized recording at the SDK level. The browser's getDisplayMedia API is the approved method for third-party recording.

Setting up Scriber with Doxy.me

Prerequisites:

Active Doxy.me account with a room (free or paid).
Scriber account (free trial or active subscription).
Modern browser (Chrome, Edge, or Safari with recent updates).

Setup (first time):

Log into your Doxy.me account.
Open your Doxy room and create a test session or start a real session.
Inside Doxy.me, look for the Scriber widget or integration button. (This appears in the session controls area.)
Click "Start Scriber" or "Enable AI note generation."
Grant browser permissions when prompted: "Allow this site to access your microphone and speakers."
Scriber is now active and capturing audio.

During the visit:

Patient is informed (verbally at the start of the call, or via Doxy's consent prompt if you have configured it) that the visit is being recorded for documentation.
Clinician and patient talk normally; Scriber listens in the background.
You do not need to do anything — Scriber is recording.

After the visit:

The visit ends.
You choose a note template (SOAP, DAP, or a specialty template).
Scriber generates the draft note (usually within 10-30 seconds).
You review the draft, make any edits, and sign it.
The audio is deleted.

That is it. The note is ready for your EHR, signature, and storage. No manual upload, no separate files, no extra steps.

Consent and patient notification

Using Scriber with Doxy.me requires clear patient consent. Here is how to handle it:

Option 1: Consent at the start of each visit (verbal) "This visit will be recorded, and an AI transcription tool will help generate your note. Is that okay?"

Patient agrees verbally. You document in the note: "Recording and AI documentation were discussed and consented to."

Option 2: Standing consent (intake form) Your intake form includes: "During telehealth visits, we use AI-assisted transcription to generate clinical notes. The recording is deleted after the note is created. Do you consent?"

Patient checks a box or signs. This consent covers all future visits (unless they withdraw it).

Option 3: Doxy.me's built-in consent If Doxy.me offers a consent prompt (some versions do), you can configure it to notify patients at the start of the call that recording is happening. This is the most automated approach.

Why this matters: In all-party consent states (see Recording consent laws by state), you must have explicit patient consent to record. Even in one-party states, transparency is best practice. Patients appreciate knowing what is happening with their data.

Troubleshooting Doxy.me + Scriber issues

"Scriber is not showing up in my Doxy.me session."

Check:

Are you using a supported browser? (Chrome, Edge, Safari — Firefox may have limitations.)
Is the Scriber integration enabled in your Scriber account settings?
Have you granted browser permissions for microphone and audio output?

If none of these help, contact Scriber support.

"Audio is choppy or the transcript is incomplete."

This usually means:

Your internet connection is unstable. Check your bandwidth.
Your browser's audio output is not being captured fully. Try refreshing the Doxy session and restarting Scriber.
If the issue persists, Scriber's support team can review the capture quality.

"The note is ready but has errors or missing information."

This is normal. The AI generates a draft, not a finished note. You review and edit before signing. If you see missing sections:

Check the original transcript in Scriber's interface. Does the transcript have the missing info? If yes, edit the note. If no, the AI may have missed it during note generation — edit the note to add it.
If the patient said something and neither the transcript nor the note captured it, it may be an audio quality issue. You can manually add the information to the note.

Remember: you are responsible for the accuracy of the note before you sign it. The AI is a tool, not a replacement for your review.

"The patient can see Scriber recording in Doxy.me. Is that a problem?"

No — in fact, it is transparent and good practice. If your Doxy.me session shows a "Recording active" indicator or the patient sees Scriber starting, that is transparency. Patients should know they are being recorded.

Audio quality and note generation

The quality of Scriber's output depends partly on audio quality:

Best practices for audio:

Ensure good internet connection. Both you and the patient should have stable, reasonably fast internet (5+ Mbps).
Use a good microphone. Built-in laptop microphones are often noisy. A headset or external USB microphone improves clarity.
Patient microphone quality. You cannot control this, but patients often have poor audio. Encourage them to use a headset if possible.
Minimize background noise. Quiet room, no loud background noise, muted notifications.

When audio is clear, Scriber's transcription is typically 95%+ accurate for clinical speech. When audio is noisy or garbled, accuracy drops. This is true for all speech-to-text systems.

If audio quality is an issue:

Ask the patient to move closer to their microphone or use a better one.
You and the patient can pause if you need to move to a quieter location.
If the audio is still problematic, the note will have gaps. You fill them in during review.

Compliance: HIPAA and state-specific

HIPAA: Doxy.me is HIPAA-compliant. Scriber is HIPAA-compliant (with BAA). Together, they are compliant. Audio is encrypted in transit, processed on BAA-covered infrastructure, and deleted after note generation.

State recording consent laws: See Recording consent laws by state for specifics. If you are in an all-party consent state, ensure you have explicit patient consent before starting the recording. One-party consent states are more permissive, but transparency is always best practice.

42 CFR Part 2 (SUD): If you are a SUD program using Doxy.me + Scriber, the same Part 2 compliance rules apply. Patient must consent to recording and AI processing. Audio is not retained. Note goes into your Part 2-compliant record system.

Multi-clinician or group practice considerations

If you are a group with multiple clinicians:

Each clinician has their own Doxy.me room and their own Scriber account integration, or one shared Doxy.me account with role-based access in Scriber.
Patient records are separated by clinician. Clinician A's notes are in Clinician A's Scriber workspace; Clinician B's notes are in their workspace.
Audit logging shows which clinician generated each note. This is important for supervision, QA, and compliance.

Scriber's role-based access controls ensure each clinician sees only their own patients' notes (unless you explicitly set up shared access for supervision or co-signature).

Bottom line

Scriber's integration with Doxy.me is what telehealth AI documentation should be: seamless, compliant, and invisible to the patient experience. You start the visit, Scriber listens, the patient has no idea anything unusual is happening, and when the visit ends, a draft note is ready for your review. No manual upload, no separate audio files, no extra steps. Patient consents, you have a compliant workflow, and you save 10-15 minutes per visit on documentation.

For clinicians using Doxy.me, this integration is the differentiator that makes AI scribing practical.

Doxy.me and AI Scribes: Best Practices | Scriber Wiki