Reviewer
Scan synthetic claim text for safety issues, route it with the on-account Workers AI classifier, extract the four IDR decision factors, then assemble a source-cited, sealed readiness dossier. All samples below are synthetic — names, dates, and figures are fictional.
Step 1 — Source text
Step 2 — Route claim type
Uses the scan result from Step 1. A tiny open-weight model
(@cf/meta/llama-3.2-1b-instruct) running on this account's
own Workers AI classifies the claim type — no third party sees the text.
Step 3 — Extract factors
Uses the scan result from Step 1. The four factors are extracted and anchored to source spans.
Step 4 — Assemble & seal dossier
Assembles the extracted factors into a readiness dossier with a SHA-256 seal. Click a source-span anchor to jump to the supporting text.
Source text with spans
Spans returned by the API are highlighted below once extraction runs.