Product
How is BrainScribe different from an ambient consultation-scribe tool?
An ambient consultation-scribe tool listens to a consultation and produces a clinical note. The unit of work is the consultation. BrainScribe's unit of work is the assessment. Scales, performance-test scores, observations, informant reports, and interview transcripts live on one record, and the report draft is composed from that record. Source-to-draft trace is the differentiator: every sentence in the draft cites the data row it was written from.
The two product categories solve adjacent problems. A specialist who runs both consultation-heavy and assessment-heavy days may get value from both.
Will BrainScribe include video sessions, recording and scheduling — or is it just the report-drafting layer?
The closed-beta vision is one workspace for the whole assessment encounter: appointment scheduling, electronic consent for telehealth and recording, browser-based video session, audio-only or audio + video recording, identified-speaker transcript, scale and test administration, observations, and the source-traceable report draft — all on the same assessment record. The video and recording layer follows a publisher-licence-style integration pattern; that relationship is on the roadmap, not active today.
What is honest to say now: the report-drafting layer, scoring of scales, and the assessment-record model are operating in the engine. Scheduling, browser video, recording and live transcript routing are planned for the closed beta and will land in cohort-staged releases. The integrated session page describes the full design and labels the planned components plainly.
Will my notes be used to train AI models?
No. Patient and assessment data is not used to train any automated model, BrainScribe's or a vendor's. The platform processes content through Claude under a commercial agreement that excludes training. Identifiers are separated from clinical content before any automated processing occurs. The same commitment is in the privacy policy and the terms of use.
Does BrainScribe replace clinical judgement?
No. BrainScribe drafts. The clinician decides. Every output is a draft for the practitioner's review — every section, every recommendation, every cited interpretation. Drafts arrive marked REVIEW_REQUIRED until the clinician approves them. Exported reports carry an audit-trail footer naming the interpretation library version, the fragment count, and the review state at sign-off.
Is BrainScribe a medical device?
No. BrainScribe is not a diagnostic instrument and is not classified as a medical device under the Therapeutic Goods Act 1989 (Cth) or the Therapeutic Goods (Medical Devices) Regulations 2002. It is a clinical workspace that supports the practitioner's documentation, scoring, and report-drafting workflow. The diagnostic opinion is the clinician's.
Clinical
Which scales does BrainScribe support today?
Twenty-four reference scales are live in the closed beta — five fully published with per-scale recommendation libraries (PHQ-9, GAD-7, NPI-Q, ESS, IQCODE-SF), and nineteen further scales in active validation. Coverage spans mood and anxiety, cognition and informant report, ADHD and developmental, trauma, substance use, function and capacity, child and family, and sleep, pain and wellbeing.
The roadmap targets 138 scales by general availability. The scales page lists what's grouped where; the changelog tracks what shipped last.
Which performance tests are supported?
Performance tests are a separate module from scales. The closed beta covers raw-to-scaled conversion for cognitive instruments commonly used in Australian neuropsychology practice — Wechsler index scores (WAIS-IV, WISC-V, WMS-IV), RBANS, NEPSY-II subtests, HVLT-R, BVMT-R, TOMM, NART, Trail Making A/B, Stroop, COWAT/FAS, Grooved Pegboard, and CVLT-3 are scoped for the v1 release. Norms are stratified by age, education, and — where the published data permits — sex. Australian-sourced datasets are flagged in the catalog. See the tests page for the working catalog.
How does the source-to-draft trace work?
Every drafted sentence in a report carries a citation back to the data row that produced it — a scale score, a transcript span, a test subscale, or an informant response. The citation is rendered as a superscript in the editor and as a footnote in the exported PDF. Clicking the citation opens the source row in a side panel.
The mechanism is structural, not statistical. Drafts are composed from fragments that are bound to source records at write time. A sentence cannot land in a draft without a source attached — that is the engine's invariant, not a quality estimate.
Can I edit the draft, or is it locked?
The draft is fully editable. Every section is rich text. The clinician rewrites sentences, deletes paragraphs, adds custom sections, and reorders the document — the way a Word draft works. Edits are tracked at the section level for the audit trail, not at the keystroke level. When the report is signed, the final text is what the clinician wrote, not what the engine drafted.
Trust & compliance
Where are my records stored?
Records are held within Australian-standard infrastructure, encrypted in storage and in transit. Patient identifiers are separated from clinical content before any automated processing. Backup and disaster-recovery copies remain within Australian jurisdiction. The security page covers the architecture in more detail; the privacy policy covers the legal framing.
Who can access my clinical data?
The practitioner who owns the account, and — for Practice tier — members of that practice on a least-privilege basis. BrainScribe staff do not access identifiable patient data except where a support ticket the practitioner has raised requires it, and only with the practitioner's knowledge. Every access event is logged. The platform has no advertiser, broker, or third-party-research data path.
What happens to my data if I cancel?
On cancellation, patient and assessment data is deleted within thirty days unless the practitioner requests earlier deletion or full export. Account metadata is retained for seven years to meet medico-legal record-keeping expectations under Australian health-records legislation. Export is available in DOCX and PDF for individual reports, and as a structured archive for the full account. The privacy policy has the formal retention table.
Is BrainScribe AHPRA-compliant?
BrainScribe is built around AHPRA-aligned access controls — account registration requires an AHPRA number, the responsibility acknowledgement at onboarding cites the Code of Conduct, and every exported report carries the practitioner's name and AHPRA number on the cover sheet. The platform supports compliance with the practitioner's obligations; it does not discharge them. The clinician remains the responsible party for every clinical opinion that leaves the platform.
Beta & access
Who can join the closed beta?
AHPRA-registered specialists practising in Australia. Closed beta is open to clinical neuropsychology, clinical psychology, educational and developmental psychology, forensic psychology, and psychiatry. Other specialties are welcome to register interest for the general-availability waitlist. The current cohort is capped to keep onboarding hands-on; new spots open as existing participants complete their first assessments.
When does the closed beta end?
The beta runs through to general availability in late 2026. The exact GA date is set by clinical-validation completion, not by a fixed launch deadline. Beta participants keep founding-member pricing and access continuity into the GA window — there is no forced re-onboarding. The pricing page covers the stage-by-stage rate structure.