Best AI Medical Scribe for Therapists in 2026: 5 Tools, Honestly Reviewed

The phrase “best AI medical scribe for therapists” returns a lot of search results, almost none of which were written by a clinician who actually does talk therapy. Most “best of” lists are aggregations of hospital-grade documentation tools that happen to mention therapists in a footnote. A general tool-of-the-year list is not useful for someone doing trauma work, addiction recovery, or couples therapy.

This article is different. The shortlist below is restricted to tools built for licensed mental health practitioners. It comes from a multi-clinician group practice that has been through this purchase decision twice in the past three years. The five tools below are the ones I would put on a shortlist today if I were starting that process again.

The list is short by design. The market is crowded: fifty-plus tools, half claiming therapist support. Most are general medical tools with a “therapy” checkbox. The closing section names them and explains why they did not make the cut.

What makes a tool the best AI medical scribe for therapists

Before the reviews, three filters that everything below has to pass:

Modality awareness. Therapy is not generic clinical work, and modality language matters: cognitive restructuring is CBT, defusion is ACT, distress tolerance is DBT. A note that does not know which technique belongs to which framework will feel wrong every time you read it. The right AI medical scribe for therapists adapts its language to the modality you actually practise.

The psychotherapy-notes versus progress-notes split. Under HIPAA, psychotherapy notes (your hypotheses, formulation, countertransference) are defined at 45 CFR § 164.501 and receive heightened protection: specific authorisation is required for disclosure under § 164.508(a)(2), and they are excluded from client right of access under § 164.524(a)(1)(i). They have to be stored separately from the clinical record. Most generic tools treat all notes as a single store. Therapy-specific platforms build the split in by default.

Where the tool sits in the session. Most AI documentation tools work after the session, generating notes from a transcript. A few work during the session, surfacing intervention prompts and tracking modality cues in real time. Both can produce good documentation, but they solve different problems. Decide which kind of AI medical scribe for therapists you want before reading the reviews.

The five tools below all pass the first two filters. They differ on the third, and the reviews call that out explicitly.

1. Emosapien

Best for: therapists who want an active AI co-therapist plus the rest of the practice workflow, not just post-session documentation.

Emosapien positions itself as an “AI co-therapy platform” rather than a transcription tool, and the distinction is real. It runs during the session, surfacing modality-aligned intervention prompts (cognitive restructuring for CBT, defusion for ACT, distress tolerance for DBT) and drafts a structured progress note in SOAP, DAP, BIRP, or GIRP afterwards with treatment-plan goals already linked. The note feels written from inside the session, not reverse-engineered from a transcript.

Scope is the other distinguishing thing. Emosapien covers the full therapy workflow: intake, in-session co-therapy, documentation, and between-session client engagement (check-ins, AI-assisted journaling, homework). For practices that would otherwise pay for three or four tools (an EHR, separate documentation, an engagement app, and analytics), this is what changes.

The free tier is also the most generous on this list: ongoing, no credit card, no countdown. Most alternatives run time-limited trials, which makes evaluation on real clients harder.

What stands out:

• Active in-session AI co-therapy. Live intervention prompts in your modality, configurable per-session so cautious clinicians can run silent and turn prompts on only for difficult cases.
• End-to-end workflow. Intake → in-session → documentation → between-session engagement, in one platform. The Engagement Agent keeps therapeutic momentum going between weekly sessions.
• Modality-driven notes. CBT/ACT/DBT packs change the language of the Assessment section, not just the section structure.
• Compliance posture for therapy specifically. Separate vault for psychotherapy notes (the HIPAA distinction most tools miss), 42 CFR Part 2 support for SUD records, and regulator-aligned defaults across major markets (APA, AHPRA / PsyBA).

Where it could be stronger:

• Direct EHR integrations are still building out: SimplePractice is live today, TherapyNotes and Jane are in beta, and Halaxy, Power Diary, and Practice Better are on the connector roadmap.
• The breadth of the platform means a learning curve if you only want a documentation tool and nothing else.

Best fit: solo and group therapy practices that want documentation and active clinical support across the client lifecycle, in one tool. See the pricing page for plan details.

2. Mentalyc

Best for: US-based mental health practices that want a mature, EHR-integrated note generator with a long track record.

Mentalyc was one of the first AI documentation tools built specifically for mental health, and it shows in the polish of the product. It generates SOAP, DAP, BIRP, and GIRP notes from session transcripts, integrates well with the major US therapy EHRs, and has a reputation for reliable, low-fuss output. Where many tools added “therapy” as a feature, Mentalyc started there.

The Assessment section reads cleanly, even if it tends towards generic clinical language rather than modality-specific framing. For clinicians who do not want their notes to “sound” like CBT or ACT, this is a feature; for those who want their modality voice in the chart, it is a limitation.

What stands out:

• Long-running US-market focus, with deep integration into the EHRs US therapists actually use.
• Strong support for the four therapy-standard note formats.
• HIPAA-aligned with a BAA on paid plans.

Where it could be stronger:

• The notes are generated post-session. The tool does not participate during the session.
• Modality-specific drafting is limited compared to tools that have explicit CBT/ACT/DBT packs.
• The free trial is time-limited rather than the ongoing free tier Emosapien offers, which makes evaluation on real clients harder.

Best fit: US-based solo or group practices on a major therapy EHR, who want a proven, low-fuss documentation engine and do not need active in-session features.

3. Upheal

Best for: therapists who want strong documentation plus session-level analytics.

Upheal generates therapy-focused notes (including SOAP, DAP, BIRP) and adds an analytics layer that surfaces patterns across sessions: alliance trajectory, themes, attendance, and progress signals. For practice owners who want a dashboard of caseload-wide indicators, Upheal is one of the few tools that includes this in the core product rather than as an add-on.

Like Mentalyc, Upheal works after the session, but the analytics are useful between sessions. The therapist can see how a client’s affect has trended over a treatment course, or whether the same theme keeps surfacing without resolution. That is genuinely useful clinical information, even if it does not change what happens in the next session itself.

What stands out:

• Built-in session-level analytics across the therapist’s caseload.
• Solid SOAP and DAP note quality.
• Active US and European therapist user base.

Where it could be stronger:

• The Assessment section, like Mentalyc’s, leans generic clinical rather than modality-specific.
• The tool does not participate during the session.
• The analytics dashboard is useful but can become noise if the practice is not actively using the data to inform supervision.

Best fit: group practices and supervision-heavy environments where caseload analytics inform clinical decisions, plus solo therapists who like data-rich reflection.

4. Blueprint

Best for: measurement-based-care practices where outcomes data drives the work, and notes are secondary.

Blueprint started as a measurement-based-care platform and added AI note generation more recently. The orientation shows. Outcome assessments (PHQ-9, GAD-7, custom scales) are first-class citizens, and the notes wrap around them rather than the other way around. For practices that already use measurement-based care, Blueprint is the most complete option on this list.

For practices where notes are the primary output and outcomes data is secondary, Blueprint can feel inverted. The note generation is competent but lighter than Mentalyc’s or Upheal’s, and the modality-aware drafting is the weakest of the five tools reviewed here.

What stands out:

• Measurement-based-care infrastructure built in (outcome assessments, scoring, trending).
• Strong on goal-tracking and outcome-anchored progress notes (GIRP-style).
• Good integration with major therapy EHRs.

Where it could be stronger:

• Notes are competent but not as developed as the documentation-first tools.
• The active in-session features are limited.
• The tool is more expensive than the alternatives on a per-clinician basis, especially if you only want the documentation side.

Best fit: practices that already run, or want to run, a measurement-based-care programme. Less of a fit for practices that just want fast, modality-aware documentation.

5. Supanote

Best for: solo therapists and small practices who want a focused, lightweight tool with no extra features.

Supanote does one job and does it well: it generates clean therapy notes from session audio, in the formats therapists actually use, with a minimum of administrative friction. The interface is fast, the export options are sensible, and the price point sits below most of the alternatives. For clinicians who want a documentation tool and nothing else, Supanote is a serious contender.

The tradeoff is that you are buying exactly that. There is no analytics layer, no engagement agent, no measurement-based-care plumbing, and no active in-session participation. If those are features you would like to ignore, that is a feature, not a bug.

What stands out:

• Lean product focus. The tool does notes, well, and stops there.
• A focused product scoped for solo practitioners, with low operational overhead.
• Fast onboarding and a low learning curve.

Where it could be stronger:

• No active in-session features.
• Limited modality awareness compared to Emosapien.
• Smaller integration ecosystem than Mentalyc or Upheal.

Best fit: solo therapists who want a no-frills tool with a narrow feature set, or practices that need a documentation layer for the few clinicians not yet ready for a fuller tool.

Quick-reference comparison

The dimensions therapists actually use to shortlist a therapy-specific documentation tool, scored across the five tools. Detail in each review above.

Capability	Emosapien	Mentalyc	Upheal	Blueprint	Supanote
Active AI co-therapy: live in-session prompts	✓	✗	✗	✗	✗
Modality packs (CBT / ACT / DBT) drive the note	✓	◐	◐	✗	✗
Therapy note formats (SOAP / DAP / BIRP / GIRP)	✓	✓	◐	◐	◐
Custom note templates	✓	◐	◐	✓	◐
Between-session client engagement (check-ins, journaling, homework)	✓	✗	✗	◐	✗
Treatment-plan continuity across sessions	✓	◐	◐	✓	✗
Separate psychotherapy-notes vault (HIPAA distinction)	✓	✗	✗	✗	✗
Multi-jurisdictional regulator-aligned defaults (APA / AHPRA / PsyBA)	✓	✗	✗	✗	✗
Ongoing free tier: no credit card, no countdown	✓	✗	✗	✗	✗
Workflow scope (intake → in-session → notes → engagement)	Full	Notes only	Notes + analytics	Notes + outcomes	Notes only

Legend: ✓ fully supported, ◐ partially supported, ✗ not available

For a deeper format-by-format look at the note formats these tools generate, the clinical documentation hub covers SOAP, DAP, BIRP, and GIRP with worked examples.

Why this list excludes Heidi, Nuance DAX, Suki, Abridge, and Freed

The general medical documentation tools are excellent for their job that do not solve the right problem for therapy. Heidi is widely used in Australian primary care; Nuance DAX has deep enterprise hospital integration; Suki, Abridge, and Freed are popular among physicians and specialty clinicians. They are HIPAA-aligned and well-built; the issue is fit. An AI medical scribe for therapists has a different shape than a medical scribe pointed at therapists.

What they are not is built for therapy. The gaps:

Note shape. Their default SOAP note is a medical SOAP. The Subjective is symptoms; the Objective is vitals; the Assessment is differential diagnosis; the Plan is medications and follow-up labs. A therapy SOAP needs none of those headings to be filled the same way. You can configure these tools to skip the medical fields, but you are working against the grain of the product.

Modality voice. None of the general medical tools carry CBT, ACT, or DBT framing into the Assessment section. They produce a clinically competent paraphrase of the conversation. A therapist’s note has to do more than that. It has to record the intervention used, the client’s response, and the modality logic that connects them.

The notes split. These tools have one notes store. They do not natively distinguish between psychotherapy notes (the protected process notes under HIPAA) and progress notes (the clinical record). For a therapist, that is not a polish issue; it is a compliance issue. See our guide to HIPAA-compliant therapy notes for the specifics.

If you do anything other than talk therapy, the bigger names are the right place to start. If you do talk therapy, they are not.

How to evaluate an AI medical scribe for therapists for your practice

Three suggestions:

1. Run real sessions, not demo sessions. Vendor demos are scripted to make the product look good. The only honest test is whether the tool produces a note you would sign for one of your actual clients. If the tool offers a free tier (Emosapien) or a free trial (the others), use it on a session with permission from a colleague playing client, then on a real session with informed client consent.
2. Read the Assessment, not the Subjective. The Subjective and Objective will look fine on every tool. The Assessment is where the differences become obvious. If the Assessment reads like a paraphrase of the transcript, the tool is a transcription engine; if it reads like clinical work, it is a therapy tool.
3. Check the BAA before the price. If you handle PHI under HIPAA, the BAA is non-negotiable, and the BAA terms vary across these tools. Read them. The American Psychological Association has practice resources at apa.org/practice on documentation and HIPAA compliance that are worth reviewing alongside any vendor’s BAA.

Pick the tool that gives you the best note for your specific caseload, not the one with the best demo. If you also want active in-session co-therapy, start free with Emosapien and decide for yourself.

This article reviews tools as of March 2026. Pricing, features, and integrations change quickly in this market. Confirm current capabilities directly with each vendor before purchasing. Educational content for licensed mental health practitioners; not legal, clinical, or compliance advice.