How the reading works
By ‘ADHD-relevant’ we mean school-report language that research and clinical literature often considers relevant when looking at attention, activity, effort, organisation, emotional regulation and related patterns. This is not a diagnosis.
What the reading does
The reading does five things, all of them to the records:
- It groups different wordings that describe the same kind of classroom experience.
- It notes how strongly each comment is worded, in the teacher’s own terms.
- It counts how many separate years and separate voices describe something — and never counts the same teacher twice as two voices.
- It sets earlier reports beside later ones and describes how the records sit across the school years — and stops there.
- It also reads for what a detailed report does not mention.
The output is built from fixed sentences around the teachers’ own words, quoted exactly — never a written-for-you narrative.
What WholeRecord is — and is not
WholeRecord is a systematic reading of school reports for ADHD-relevant patterns.
WholeRecord is not a medical device, not a diagnostic tool, not a screening tool, not an assessment, and not a substitute for clinical judgment.
What school reports can and cannot show
School reports do not show everything. They do not show internal experience, life outside school, or the effort behind a composed surface; and they only contain what teachers chose to write.
A record can show the coping working rather than the cost underneath. A quiet, well-marked record is therefore read as what it is — a record in which these patterns were not described — never as proof there was nothing to describe.
Absence in the records is not absence in the person.
How AI is used
WholeRecord uses AI in two places: to read the words on your uploaded reports, and to sort those words against a fixed list of patterns that research links to ADHD. The reading itself is built from fixed wording we prepared in advance, with your teachers' words quoted in, as they appear in your uploaded reports — the AI does not write the explanations. WholeRecord reads records; it cannot diagnose.
The AI reading runs on Amazon Web Services in Europe. The AI models are made by Anthropic and run inside Amazon’s own European service, called Bedrock. Your reports are processed only for the moment it takes to do the reading and no copy is kept afterwards; under Amazon’s terms, Anthropic is not given the words from your reports. Who processes what is set out in full in the privacy notice.
What the testing does and does not show
WholeRecord structures what school reports already say. Its testing so far checks consistency — that the method reads the same records the same way — which is different from clinical accuracy. It has not been validated as a clinical instrument, and it makes no accuracy claim.
Where it sits
WholeRecord does not direct anyone toward or away from clinical assessment. Some people read it before one, some after, some on its own. The reading is not clinical advice.
References
Every reference on this page is checked against its source before it is published; no number from the research is presented as a claim about this product.