JotPsych

The scribe is the origin and still the anchor: the vendor's own framing invites practices to start with the scribe and add billing, credentialing and a website later. The behavioral health note structures and DSM-5-TR aware coding are model work rather than templating.

One reason to revisit rather than a deduction now: the platform expansion into billing, credentialing and e prescribing is the kind of growth that shifts a moat from the model to the switching costs of an all in one system, and this grade should be re examined once that suite matures.

Conventional draft and review with two useful additions. A session transcript is available for review before the note is finalised, which lets a clinician check the generated note against what was actually said rather than against memory, and editing is done by plain English instruction with the system learning preferences over time. Because the base workflow is copy and paste into the record, a human necessarily handles every note. Held at B because no acceptance rate, edit burden figure or confidence threshold is published.

No accuracy figure, model card, named models or evaluation methodology located. The performance claims are speed and superlative rather than correctness: a note in under 60 seconds, 90 percent of notetaking time saved, and better than human scribing, that last phrase being an unfalsifiable comparison with no defined human baseline. It belongs with the other accuracy overclaims this index tracks.

Nothing identifies any party in the chain: no model or model family, no foundation model provider, no hosting arrangement and no sub processor list was located in two passes, and nothing was found on whether customer sessions or notes are used to train models. The only retention statement located is that your data does not hang around, which names no period, no deletion mechanism and no scope, and it is worth noting as a form rather than dismissing as vague.

Casual reassurance is a distinct disclosure failure from silence: it leaves a reader with the impression that the question has been answered while committing the vendor to nothing, and it is harder to follow up on than an acknowledged gap because a therapist reading it may believe they already have the answer.

In a segment where the input is recorded psychotherapy, the two things a buyer needs are a retention period with a deletion mechanism and an explicit training position, and a peer in this same category publishes both in plain terms. Ask for a stated retention period, the deletion mechanism, the training position in contract language, and a sub processor list.

Adoption is real and named, with more than 10,000 behavioral health practices claimed and several identified by name including psychiatric groups and a TMS practice, but adoption is not benefit. No study, controlled evaluation, accuracy benchmark or third party performance rating was located. Graded C on the standing precedent that scale of use does not substitute for evidence.

This is the weighted axis for behavioral documentation because the input is recorded psychotherapy, and the disclosure here is thin. The only retention statement located is that your data does not hang around, which names no period, no deletion mechanism and no scope, and nothing was found on whether customer sessions or notes are used to train models.

Compare Mentalyc, graded PHI-A in this index for an explicit does not train and does not retain commitment backed by SOC 2 Type II and PHIPA. Same segment, same sensitivity of input, materially weaker disclosure.

HIPAA compliance stated consistently across product materials. Business associate agreement terms are not published for inspection, and the product is self serve with a free trial requiring no credit card, so a clinician can begin recording sessions before any agreement conversation.

A correction first: a security page does exist and the earlier pass did not reach it. What it documents is a data minimisation architecture rather than an attestation, and the architecture is genuinely better than most at this tier.

Stated controls, each specific enough to test: original recordings destroyed the moment transcription completes, identifier redaction applied to transcripts and charts, configurable retention with auto deletion windows, notes deleted on export to the record system, no backups or cold storage retained after deletion, clinical sessions never used to train or fine tune models, and a business associate agreement as standard rather than a paid addition. The stated design principle is that the less data held, the less there is to protect. For a product handling psychotherapy audio that is the right instinct and it is credited here.

What is absent is independent examination. No SOC 2 report of either type, no HITRUST certification and no ISO 27001 were located.

That absence is sharper than usual because of where the vendor has chosen to stand. Its own competitive comparison material faults a named competitor for providing less clarity on additional compliance certifications, giving SOC 2 and HITRUST as the examples. A vendor that identifies those two certifications as the relevant test, and applies it to a rival, invites the same test of itself. Note also that comparison material published by a vendor about its competitors is a self interested source and this index treats it accordingly, but the criticism is being taken at the vendor's own word rather than against it.

Ask which report is held, of which type, and covering what period.

No clearance claimed and no United States device pathway attaches. This axis does not read as an absence here, because behavioural health documentation sits inside a regulatory structure that is genuinely more demanding than the one governing a general medical scribe. It is simply not device regulation.

Three things govern rather than one. Psychotherapy notes carry protections beyond ordinary clinical records, with separate handling and disclosure requirements. Substance use disorder treatment records fall under a distinct federal confidentiality regime with its own consent rules, and this vendor supports addiction treatment templates, so that regime reaches part of its footprint. And recording a therapy session engages state consent law, which varies by state and applies to the recording itself rather than to the documentation produced from it.

The vendor engages directly with the third of those, which is what lifts this above the category baseline. It publishes client consent language for clinicians to use, covering what the tool does, that audio is deleted once the note is created, that the clinician reviews and finalises every note, and that a client may decline at any time without affecting their care. Very little in this category supplies its users with consent wording at all, and in a setting where the recording is of a therapeutic conversation that is a substantive contribution rather than a courtesy.

What is not addressed is the substance use disorder regime specifically. A practice documenting addiction treatment should establish how the product handles records falling under it, since the consent architecture there differs from ordinary health privacy consent.

On scope, the product also performs diagnostic and procedural coding including the behavioural health diagnostic classification, which is the reimbursement boundary this index tracks across the category.

Graded on incentive structure, with a genuine distinction to hold.

JotPsych describes its notes as audit proof by design, carrying the detail, time and language payers want to see, and states that they document complexity and support the code, while its billing arm charges a percentage of what it collects. That is the second instance in this wave, after Steer Health, of contingent pricing combined with documentation shaped to support billing level, and it is stated more explicitly here.

In fairness, part of it is legitimate and specific to this specialty. Psychotherapy CPT codes are time based, so a note that records session duration is meeting a real documentation requirement rather than inflating anything, and behavioral health has genuine under documentation problems.

The concern is narrower than the phrasing suggests, and it is this: when the same vendor writes the note, decides what complexity it documents, and takes a share of what the resulting claim collects, no independent party sits between the documentation and the payment. No fairness, subgroup or accent disclosure was located.

Two passes located no accuracy or error figure, no published limitations, no evaluation methodology and no warranty, indemnity or remediation commitment. The performance claims describe speed and superiority rather than correctness: a note in under sixty seconds and a large proportion of notetaking time saved both measure duration, and the claim of being better than human scribing is an unfalsifiable comparison, because no human baseline is defined, no scribing population is identified and no measurement of either side is published.

It also invites a reading the vendor may not intend, since better than a human at this task would be a strong claim about a system that has published nothing about how often it is right. The consequences of an inaccurate note in this segment are heavier than in general documentation.

A psychotherapy progress note supports a clinical formulation, is submitted for reimbursement, and is discoverable in custody, criminal and civil proceedings, so it is read later by parties the client never anticipated. Risk content is the case that most needs a published position, since a note that omits or misstates disclosed suicidal ideation, abuse or substance use carries immediate clinical and legal consequences, and nothing describes how the system handles it. Ask for a measurement with a defined baseline, for how risk content is treated, and for what the vendor commits to when a note is wrong.

The base workflow is transfer rather than integration: record, review the generated note, then copy and paste it into the record system. EHR integration exists but is offered as an ADD ON on paid plans, which is the pricing tell this index now reads deliberately, since what a vendor gates tells you what the base product lacks. For a solo behavioral health practice that may be entirely acceptable, and it does preserve a firm human gate, but a buyer should not assume structured write back is included.

Better answered than previously recorded, and the vendor's retention posture changes what this axis is protecting.

What is established: data handling is described as United States based, infrastructure is described as meeting health privacy requirements, encryption is named to the strength in transit and at rest, original audio is destroyed when transcription completes, identifiers can be stripped from transcripts and charts, retention windows are configurable with automatic deletion, and notes are deleted once exported to the record system.

That materially reduces what residency would protect. Where a vendor holds encounter audio and transcripts indefinitely, the question of which jurisdiction's servers hold them and whose authorities can reach them is the whole question. Where audio is destroyed at transcription and notes leave on export, the persistent surface is small by design.

The exposure moves rather than disappearing, and this is the point a buyer should press. It concentrates into the processing window itself, which is precisely the part left undescribed. Nothing published establishes where transcription and note generation run, whether a third party model service performs either, which provider that would be, or what that provider retains during and after the call. A commitment that the vendor keeps nothing says nothing about what a subprocessor keeps.

No hosting region is named beyond the country level, no residency option is offered, and no subprocessor list was located.

For psychotherapy audio the processing window deserves more scrutiny than it would elsewhere, because a session recording is among the most sensitive artefacts any vendor in this index handles. Ask which providers touch it, under what terms, and for how long.

A full published ladder across several products, which is rare anywhere in this index. The scribe is 150 US dollars per month for an individual and 130 per user per month for groups of 7 to 50, with custom pricing above that. The wider practice platform starts at 299 dollars a month for core software.

Most notably the contingent elements are disclosed rather than hidden: billing is charged as a percentage of collections, credentialing is priced per payer, and the website product is a one time build plus hosting. A free trial requires no credit card. Publishing that billing is contingent is exactly the disclosure this axis rewards, even though the same fact drives the governance grade down, and those two judgements are consistent rather than contradictory.

The deepest behavioral health coverage in this index, and it spans the whole specialty rather than one segment of it. Where Mentalyc and Upheal centre on talk therapy and Supanote on therapy note formats, JotPsych covers psychiatry, psychology, therapy and counselling together, with 60 or more behavioral health specific note sections across psychiatric evaluation, medication management, therapy progress notes and intake, plus templates for TMS, Spravato and polypharmacy that no general purpose scribe attempts.

Coding support includes DSM-5-TR alongside ICD-10 and CPT, which is the diagnostic standard this specialty actually uses. Longitudinal context is carried by referencing prior notes when drafting, which matters more in ongoing psychiatric care than in episodic medicine.