Overview›Build status
What's live, what's planned, where to look.
This page maps every product surface BrainScribe describes to one of three states: live in the engine, illustrative on this site (design mockup, not yet built), or planned for the closed beta. The honesty framing across the rest of the site abbreviates these. This page spells them out.
Live in the engineOperating today. Sample output is on the reports page.
Illustrative on the siteThe design surfaces on this site that show the workflow are mockups. The mechanism behind them is built or planned, but the UI shown is illustrative.
Planned for closed betaOn the build path for the closed-beta release. Not yet operating. Beta participants will see these features land in cohort-staged releases.
The scoring & interpretation engine
The engine, operating today.
The BrainScribe engine is real Python code with a versioned content library, validation gate, and rendering layer. It produces the sample reports linked from the reports page. This is what's been built before the marketing site.
| Surface | State | What that means |
|---|---|---|
| Self & informant scale scoring (24 scales) | Live | Twenty-four reference scales implemented in the engine: PHQ-9, GAD-7, NPI-Q, IQCODE-SF, ESS, CSDD, AD8, FAQ, Lawton-IADL, ZBI-12, ISI, MFIS, FBI, PSWQ, FSS, BEHAVE-AD, CBI-R, DAD, MSIS-29, MFS, MSQ, BADLS, RBDSQ, WURS-25. Each with severity bands, recommendation libraries and rendering. |
| Reference scales at v0.4.1 standard (38 scales) | Live | Thirty-eight scales now produce reports at the v0.4.1 standard (profile classifiers, evidence-summary voice, published-norm provenance). Six further scales are templated placeholders awaiting spec rebuild. |
| Per-subscale score panels with cited bands | Live | Multi-subscale scales (PCL-5 clusters, MSIS-29 physical/psych, MFIS subscales, NPI-Q domains, RPQ, ZBI-12, EMQ-R and others) render published cut-points alongside the headline severity. Bar-chart panels in HTML and PDF. |
| HTML + PDF rendering with audit footer | Live | Two output formats per scale. The audit footer carries interpretation-library version, fragment count, REVIEW_REQUIRED status at sign-off and generation timestamp into the export. Eight samples on the reports page. |
| Blank-form generation (completer-aware) | Live | Every scale produces a printable blank form, with per-item anchors where the response options vary across items. For multi-axis instruments (Yes/No + severity + distress for NPI-Q), the blank form shows all three axes per row. |
| Reliable-change (RCI) and MCID for serial review | Live | PHQ-9, GAD-7 and a growing set of supported scales produce RCI and MCID change descriptors against published references. Comparison mode for two administrations is operational. |
| Profile classifiers (e.g., mood-dominant, frontal disinhibition) | Live | Twenty-seven profile classifiers across the live scales. Same total raw score can route to different narrative profiles depending on item-level pattern. |
| Pre-render validation gate | Live | Item-total mismatches, unresolved fragments, template artefacts, and diagnostic overreach are caught before render. Pool-wide vocabulary scans run regularly. |
| Performance-test scoring (public-norm tier) | Planned | The two-tier model is described on /tests. Tier A (Trail Making, COWAT, Animal Fluency, BNT, RCFT, Stroop, MoCA, NART, Bells, Pegboard) is on the closed-beta build path against published-literature norms. |
| Performance-test integration (publisher tier) | Planned | Tier B (WAIS, WMS, CVLT, BVMT-R, HVLT-R, D-KEFS, SDMT, JLO, Warrington RMT) is score-entry — the clinician scores with their publisher's licensed materials and BrainScribe integrates the value. Workflow is on the closed-beta build path. |
The integrated session
Scheduling, consent, video, scribe — on the closed-beta path.
The integrated session described on /telehealth is the centrepiece of the closed-beta release. None of it is operating today. The video and recording layer follows a publisher-licence-style integration pattern; the partnership is on the roadmap, not in place. The session surfaces shown on the site are design illustrations.
| Surface | State | What that means |
|---|---|---|
| Appointment scheduling | Planned | Bookable slots per clinician or practice, calendar invite with the secure session link, pre-visit form and consent paperwork sent automatically on booking. |
| Calendar sync (Google / iCal / Outlook) | Planned | Read/write to the practitioner's external calendar so BrainScribe coexists with existing PMS systems rather than replacing them. |
| Electronic telehealth consent | Planned | consent.telehealth object — required before the video session can be opened. Versioned wording, captured signature, in-clinic or remote signing. |
| Electronic recording consent | Planned | consent.video_recording object — independent of telehealth consent. Recording is disabled until signed. Audio-only and audio + video share the consent with a captured mode flag. |
| Electronic transcript consent | Planned | consent.transcript object — independent of recording consent. Model-assisted processing of the recording is disabled until signed. A client may consent to record but not to transcribe. |
| Browser-based video session | Planned | Two-party video for clinician and client, no client install. Built on a third-party video infrastructure partner; the partner relationship is on the roadmap. |
| Audio-only recording (telephone or in-clinic) | Planned | Captures the consultation audio against the assessment record. Attached to the record; not a separate file the clinician has to file later. |
| Audio + video recording (telehealth) | Planned | As above, with video. Gated by the recording-consent object. Subject to the practice-level retention schedule. |
| Identified-speaker transcript (live) | Planned | Speaker identification (clinician / client / informant) with timestamps. Built on a best-in-class ASR partner evaluated on Australian clinical accents. Identifier substitution applied before any model-assisted processing. |
| Transcript-to-section routing | Planned | Captured material routes to the section of the report it belongs in (presenting concern, history, mental state, risk, formulation). Held in awaiting your review status until the clinician approves. |
The clinician-facing web application
The workspace itself — pre-launch.
The clinician-facing web application — the surface where a registered clinician logs in, builds an assessment record, drafts a report, signs and exports — does not exist as a runnable product today. The engine that powers the scoring and the drafting is real (above); the workspace around it is on the closed-beta build path.
| Surface | State | What that means |
|---|---|---|
| AHPRA-verified account creation | Planned | Account gated by AHPRA registration verification. Lapsed registration disables draft generation. Pre-beta the waitlist holds an email; there is no live login. |
| The assessment record (build, view, edit) | Planned | One versioned record per case, holding client file, scales, scores, transcript artefacts and observations. The data model is specified on /developer-specification. |
| Source-to-draft trace per sentence (UI) | Planned | Click any drafted sentence to see the score, scale item, or transcript span it was written from. The engine produces traced fragments today; the UI surface that explores them is closed-beta. |
| Section-level sign-off | Planned | REVIEW_REQUIRED on each section until the clinician approves. Audit lines record the approval. The engine writes the status; the UI signs it. |
| Integrated report drafting from the record | Planned | The composed report assembles from the whole record, not just one scale. Each drafted sentence carries its citation. The single-scale PDF is the live precedent. |
| Six report templates (neuropsych, capacity, ADHD, developmental, medico-legal, NDIS) | Planned | Each template defines its sections, the inputs each section is composed from, and the export format. Listed on the /reports page; mapped against the engine in beta. |
This site & the waitlist
The site you're reading — live, with caveats.
The marketing site exists, builds, and is what's in front of peer reviewers today. The forms post to a Netlify endpoint; the beta waitlist captures email addresses; nothing on this site grants product access.
| Surface | State | What that means |
|---|---|---|
| The site itself (23 routes) | Live | You're reading it. Static Astro build, deployable to Netlify. Every page returns 200 on the dev server and builds clean. |
| Sample report downloads (engine output) | Live | Eight real engine PDFs on the reports page. These are not mockups — they're what the scoring and rendering layer produces. |
| Site illustrations (telehealth, scheduling, consent, peek cards) | Illustrative | Design mockups of the integrated session and the assessment record. The mechanism behind them is specified; the rendered UI shown on the site is illustrative pending the closed-beta build. |
| Beta-signup form | Live | Posts to Netlify Forms. Captures email + AHPRA number + specialty. Returns a "we'll be in touch" auto-reply. Onboarding to the closed beta opens cohort-by-cohort. |
| Newsletter & changelog subscriptions | Live | Same wiring as the beta-signup. Subscribers receive the closed-beta milestone announcements. |
| Contact addresses (privacy, security, partnerships) | Live | privacy@, security@, partnerships@, press@, hello@. Acknowledgement within two business days. |
Join the closed-beta waitlist.
AHPRA-registered specialists practising in Australia. Onboarding opens cohort-by-cohort. The waitlist captures your details — it isn't a login.