AI Notes for Online & Teletherapy Sessions (Zoom, Meet, Teams) (2026)
Written by Kshitij Domadia, Founder, MyKaya
Published September 3, 2026
Teletherapy has become the default mode of practice for a significant portion of therapists. The infrastructure question — how to record a session that is already digital — turns out to be more complicated than it sounds, and the approach you use has real implications for your client, your compliance obligations, and the quality of the notes you end up with.
This article covers the practical options, their trade-offs, and the approach that works best for most online therapy contexts.
The core problem with recording teletherapy sessions
Recording a session that happens over Zoom, Meet, or Teams introduces a choice most therapists do not think about until they are already in it:
Option 1: Use the video platform's built-in recording. Most platforms allow recording with host permission. The problem: the recording is visible to clients, the audio quality is often compressed, and the recording is tied to the platform's storage and terms of service rather than your healthcare-compliant infrastructure.
Option 2: Use a bot that joins as a participant. Meeting bots (like Fathom or Otter.ai) appear in the participant list. The client sees a "Notetaker" entity join the session. As discussed in the AI scribe vs meeting bot guide, this visibility changes the dynamic of the session, particularly for clients disclosing sensitive material.
Option 3: Use a desktop audio capture tool. Tools like MyKaya run from your computer and capture system audio — the audio you hear through your headphones or speakers — without joining the call or appearing as a participant. The client's experience is completely unchanged. No new participant appears. The session looks and sounds exactly as it always has.
Why option 3 works best for teletherapy
The session stays undisturbed. Clients on a video call are already in a slightly different psychological state than in-person clients — there is a screen between you, the backdrop is different, and the sense of containment of the therapy room is modified. Adding a visible recording indicator or a bot participant further shifts that dynamic. Desktop capture eliminates this entirely.
Audio quality is typically better. A desktop scribe captures your system audio directly, which means it captures the processed audio output — what you hear — rather than the raw microphone input that many bots use. For AI transcription, this often produces more accurate results.
Compliance is cleaner. When you use a platform's built-in recording, the recording lives on Zoom's or Google's servers under their terms of service. A desktop scribe sends the audio to a healthcare-compliant cloud infrastructure under your practice's BAA or DPA — much cleaner from a compliance standpoint.
You stay in control of the workflow. With desktop capture, you run the tool from your own device, and the recording and note generation happen on your timeline, not tied to the video platform's recording export process.
What the consent conversation looks like for teletherapy
Informed consent for recording an online session requires a slightly different framing than for in-person sessions, because clients may assume the video call is being recorded anyway (most platforms offer recording features that clients know exist).
A clear, honest script:
"I use a tool on my computer to transcribe our sessions and generate my clinical notes. It captures the audio through my system — not a recording that appears on your end or shows up in the call. The audio is processed by a secure service. I never share the audio with anyone. You can opt out at any time and I'll take handwritten notes instead. Any questions before we start?"
The key points to cover: how the recording works (desktop, not call-side), where the data goes, who has access, and how to opt out.
Platform-specific considerations
Zoom: Built-in recording is visible to all participants (a recording indicator appears). Local recording to your computer is less visible but still tied to Zoom's terms. Using a desktop scribe like MyKaya captures the session audio through your system audio without using Zoom's recording feature at all.
Google Meet: Google Meet does not currently have a local recording option for individual users — recording is cloud-based and visible to all participants. Desktop audio capture is the practical alternative for Meet-based therapy.
Microsoft Teams: Teams recording is cloud-based and requires admin permission in organizational accounts. Personal Teams accounts have limited recording capabilities. Desktop capture again provides the most flexible and compliance-friendly path.
In-person sessions: Desktop audio capture works identically for in-person sessions. The device captures room audio through its microphone, without any call platform involved.
The documentation output
Whether the session is in-person or online, the AI note generated should be identical in clinical quality: SOAP, DAP, BIRP, or your format of choice, with screener scores tracked in the client profile and a draft ready for your review. The session modality should not affect the note.