Introduction
Adjudicate gives you programmatic access to a Veteran’s claims file — the corpus of documents and evidence behind a benefits appeal. Through one API you can list the appeals in a workspace, read the documents in a file (with their boundaries, receipt dates, and categories), and run full-text search across every page. And it’s fast: a claims file of thousands of pages opens in well under a second — up to 7x quicker on first paint than the platform it was rebuilt from. The same data backs the Reader UI that attorneys use day to day.
What you can build
- Pull appeal data into your own workflow — list appeals, read document metadata, fetch the structured record behind a file.
- Search a claims file — full-text across every document in an appeal, with page-level hits, without parsing PDFs yourself.
- Surface evidence — resolve a citation to the exact document and page it came from.
- Build on the data model — annotations, tags, issues, and AI-extracted issue lists all hang off the same shapes.
Two ways to integrate
| Best for | Start here | |
|---|---|---|
| SDK (JavaScript / TypeScript) | Web and Node apps — typed, batteries-included auth | SDK → JavaScript/TypeScript |
| REST API | Any language, server-to-server, scripts | API reference |
Both speak to the same endpoints and the same data. The SDK is a thin, typed wrapper over the REST API — use it if you’re in JS/TS, drop to REST anywhere else.
A safe place to build
Everything here runs against an isolated sandbox on fully synthetic data — fabricated Veterans, invented documents, in a dedicated cloud account with no path to any real record. So you can explore the API, run the quickstart, and wire up an integration freely, with no PII in play. (How that isolation works: Data and the Platform overview.)
Next
- Quickstart — from zero to listing appeals in about five minutes.
- Then the SDK or the API reference, depending on your stack.
- For how the system works underneath, see the Platform overview.