The clinical translation gap
Teachers without specialist training are asked to interpret psychology, speech, and OT reports and turn them into classroom adjustments. PLPs vary widely in quality — not from lack of care, but from lack of time.
Built for Australian schools · NCCD-aligned
Personalised Learning Plans, drafted from the clinical evidence you already have.
Kindred reads psychology, speech, and OT reports and drafts a parent-ready PLP your Inclusion Leader reviews and approves — instead of writing it from scratch.
Hours
per PLP spent reading clinical assessments
$25K
per student in NCCD loading at stake
50+
PLPs a teacher may juggle each year
The problem
Clinical reports are written for clinicians. PLPs need to work for teachers and read for parents.
Someone — usually the Inclusion Leader — has to read dense neuropsych, speech pathology, and OT reports, extract what matters in the classroom, translate it into plain language, and write a plan that's useful to a busy teacher and readable by a worried parent. Then do it fifty more times before the next census.
Existing tools are filing cabinets
Junipa, ezNCCD, Tes, Inspire — they store documents and run dashboards. None of them read the clinical evidence and draft the plan. That work still lands on one person.
What Kindred does
Three jobs, one workflow: drafts, cites, evolves.
The PLP isn't a one-shot document — it lives for a school year, gets reviewed, and passes to next year's teacher. Kindred treats it that way.
01
Drafts
Reads the clinical assessments. Generates a parent-facing PLP in your school's template, plus an audit companion. Turns clinical jargon into language educators can use.
02
Cites
Every claim carries a [source:page] citation. The validator strips uncited claims before export. The audit companion preserves the evidence trail — the parent PLP reads cleanly.
03
Evolves
PLPs are living documents. Each new version carries the IL's edits forward. New evidence can be added. The teacher handoff field captures year-end summaries for the incoming teacher.
How it works
Four steps from a folder of assessments to a parent-ready PLP.
Step 1
Upload the clinical evidence
Drop in the neuropsych report, speech pathology assessment, OT review — whatever's in the student file. Kindred parses PDFs, DOCX, and scanned documents.
Step 2
Edge de-identification
Before anything leaves your browser, names, dates of birth, and identifiers are replaced with placeholder tokens. The token map stays on your device. The AI never sees the student.
Step 3
AI drafts with citations
Kindred drafts each PLP section — strengths, adjustments, goals — and tags every clinical claim with a [source:page] citation. Uncited claims are stripped before export.
Step 4
Inclusion Leader reviews, exports, evolves
Your IL edits the draft directly. Approved PLPs export as the parent-ready .docx plus an audit companion preserving the evidence trail. New evidence joins the next version.
Privacy by design
No identifiable student data leaves your school.
The honest version of the privacy story: identifiers are stripped locally before any AI call. The model drafts against tokens it can't map back to a real student. That's a different category of risk than pasting reports into a general-purpose AI chatbot — which, today, is the status quo Kindred replaces.
Identifiable data stays at the school
Names, DOBs, and other identifiers are tokenised in the browser before any request leaves your network. The token map never leaves your device.
The model only sees placeholders
The LLM works with {{STUDENT_FIRST}}, {{DOB}}, and similar tokens. De-tokenisation happens locally, after the draft returns.
Citation discipline, by design
Every clinical claim must carry a [source:page] tag. The validator strips uncited sentences before export — no fabrication survives the pipeline.
Pseudonymous audit log
Every action — generation, edit, version, download — is logged with HMAC pseudonymous IDs. Ingestible into your existing SIEM.
Australian-built, NCCD-aligned
Designed against the Victorian DET position on AI, the ST4S AI module, and the National AI in Schools Framework. No US-EdTech assumptions.
Per-school login
Each school gets its own login and its own template. Student data isn't pooled across schools.
How to get Kindred
Currently in pilot with one Australian primary school. Taking expressions of interest for Term 3, 2026.
Web-based, no install
Kindred runs in the browser. Your Inclusion Leader signs in, uploads the student's clinical evidence, and reviews the draft. No on-site server, no IT-team install.
Per-school onboarding
Each school gets its own login and its own .docx template — the draft comes back already formatted the way your school writes PLPs. We help you mark up the template once, then it's reusable across every student.
Pilot terms, written upfront
Pilot 1 is gift-priced in exchange for written feedback and a reference. Beyond the pilot, pricing is per-user and validated with the pilot school before public launch. No long-term lock-in.
If you're an Inclusion Leader, principal, or business manager at an Australian school and you'd like to talk about a pilot or an early look — the contact form below comes straight to the founder.
Get in touch
Tell us about your school.
Drop a note and Sam (the founder) will reply personally — usually within a day. If you've got a clinical assessment you'd like to see Kindred draft against, just say so and we'll set up a demo.
- Based in
- Melbourne, Australia
- Existing customer?
- Sign in to the app →