Upload patient data

Bring your own data against a template. Structured files run straight through the rules engine to a go/hold verdict; free-text notes are read by the model services first (they extract + classify candidate facts). Use synthetic data only until IG sign-off (INV-7). Verdicts also appear on the Console.

Smart ingest — drop any files

Pick the pathway, then drop one or many files of any mix — CSV, Excel, JSON, text notes/letters, scanned PDFs. Each is auto-extracted, cleansed and mapped to this pathway's facts, run through the engine, and shown with a provisional verdict + what it mapped for a clinician to confirm. Scanned PDFs go via OCR.

try:ward-list.csvscanned-referral.pdfclinic-letter.txt
Advanced: exact-format uploads (matching columns / single JSON / note) →

Batch upload — ward-list (CSV / Excel)

Upload a spreadsheet where each row is a patient and the columns are the pathway's facts (a name column + one column per fact key). Each row runs through the engine and a verdict appears live. Empty cells become awaiting. Download a sample ward-list to see the columns.

sample: surgical-readiness/ward-list.csv

1 · Upload patient data (structured)

A JSON file of coded facts for one patient (matching a template). It runs straight through the rules engine and appears below with its live go/hold verdict. Download a sample or blank template to fill in with your hospital's data.

2 · Upload / paste a clinical note (AI extracts)

The model services read free text: gliner extracts candidate facts, bert flags compliance gaps. These are advisory for a human to confirm — never the verdict (INV-1).

3 · My uploaded patients (0)

None yet — upload a patient JSON above.