Therapy & clinical
Therapy and clinical session transcription: HIPAA, SOAP notes, and the 2026 clinician playbook
Therapy session transcription, SOAP notes from audio, HIPAA-compliant transcription, BAA, clinical session intelligence, medical scribing — therapy and clinical playbook 2026.
Clinical and therapy session transcription is a regulated category
Therapy sessions, medical consultations, and any clinical conversation that touches Protected Health Information (PHI) under HIPAA cannot be transcribed with consumer tools. The vendor must sign a Business Associate Agreement (BAA), the data path must be HIPAA-eligible end-to-end (storage, transit, processing, and any subprocessors), and access controls must be documented. Free tier consumer tools rarely qualify; the few BAA-eligible transcription tools are typically on Business or Enterprise plans.
This article walks through the 2026 clinician playbook for transcribing sessions: which tools have BAAs, how to structure the workflow for SOAP note generation, what consent looks like, and the regulatory minimums you should not skip.
HIPAA-compliant transcription tools with BAAs
| Tool | BAA tier | Pricing | Diarization |
|---|---|---|---|
| Otter Business | Business plan | $30/user/month | Yes |
| Sonix Enterprise | Enterprise | Custom | Yes |
| Rev Healthcare | Healthcare plan | Custom + per-min | Yes |
| AWS Transcribe Medical | BAA available | $0.075/min (medical) | Yes |
| Google Cloud STT | BAA available | Per-min | Yes |
| Microsoft Azure Speech | BAA available | Per-min | Yes |
| DAX Copilot (Nuance) | BAA + EHR integration | Enterprise | Yes (clinical-tuned) |
| Augmedix | BAA + scribing service | Enterprise | Yes |
For solo practitioners and small clinics, Otter Business is the most-accessible BAA-eligible option. For hospital systems and large clinics, DAX Copilot (Nuance, owned by Microsoft) and Augmedix are the dominant clinical scribing solutions, with deep EHR integration. AWS Transcribe Medical is the API-first option for clinics building custom workflows. For "HIPAA-compliant transcription" specifically, all the above are workable; pick by integration with your existing EHR and scale.
SOAP notes from audio
SOAP notes (Subjective, Objective, Assessment, Plan) are the standard clinical note format in most US healthcare. Generating SOAP notes from session audio is a two-step process: (1) transcribe the session, (2) restructure the transcript into SOAP format. The restructuring step has historically been manual; in 2026, AI scribing tools (DAX Copilot, Suki, Abridge, Heidi, Freed) automate this end-to-end.
- 01Record the session with explicit patient consent (verbal recorded + written form)
- 02Upload to a HIPAA-compliant transcription service with BAA
- 03Transcribe with diarization (clinician + patient labelled)
- 04AI scribing layer extracts subjective complaints, objective findings, assessment, plan
- 05Clinician reviews and edits the SOAP draft (still the responsibility of the clinician, not the AI)
- 06Save final SOAP note to EHR via API or copy/paste
For solo therapists doing weekly progress notes, AI scribing tools dramatically reduce documentation time (industry estimates: 1-2 hours/day saved per clinician). For hospital systems, the same tools integrate with Epic, Cerner, Athena, and other major EHRs.
Consent patterns — what HIPAA actually requires
HIPAA itself does not require explicit patient consent for transcription IF the transcription is part of treatment, payment, or healthcare operations. However, ethical practice (and many state laws) requires explicit consent for recording sessions. The recommended pattern: (1) explicit verbal consent at session start, recorded as part of the session audio, (2) written consent in the patient intake paperwork that names the specific transcription service, (3) opt-out availability — patients who decline are not recorded.
For therapy specifically, the therapeutic relationship may be affected by recording — some patients become guarded knowing they are being recorded. Solo therapists in private practice should weigh the documentation efficiency gains against the potential clinical impact. For hospital settings where AI scribing is becoming standard, this trade-off has typically been resolved in favor of efficiency.
Closing: clinical transcription is a regulated, high-stakes category
For any clinician transcribing patient interactions in 2026, the non-negotiables are: HIPAA-compliant tool with signed BAA, explicit patient consent, encrypted storage and transit, and clinician review of any AI-generated documentation. The tools have matured significantly — solo practitioners can use Otter Business or one of the AI scribing tools; hospital systems use DAX Copilot, Augmedix, or similar. The cost is small relative to the documentation time saved.
Keep reading
Speaker Identification
The Speaker 1 problem: why every transcription tool fumbles who said what
9 min →
Audio to Text
Audio to text in 2026: a guide that actually accounts for accuracy, speakers, and privacy
10 min →
Video to Text
Video to text: how to convert video to clean, usable transcripts without losing context
9 min →