Getting started

How processing works

The steps a document goes through after upload, and how to follow its progress.

Every uploaded document runs through the same pipeline. The Documents tab shows which step each file is on, so you can tell the difference between something still working and something that needs your attention.

The steps

• OCR — reads the text off each page, including scanned and photographed records
• Classification — works out what each document is, and splits combined files into their separate records
• Entity extraction — pulls out dates, providers, diagnoses, treatments, and charges
• Chunking and embedding — indexes the text so it can be searched and cited later
• Timeline — places every event on the case timeline
• Intelligence — assembles the case profile and cross-document findings

How long it takes

It depends on page count far more than file count. A short office note is quick; a several-hundred-page hospital record takes longer. You can leave the page and come back. Processing carries on without the browser open.

What fills in as it goes

The case profile fills in as documents finish, so you'll see facts appear before the whole batch is done. The medical chronology and the expense summary build themselves once documents are processed. The demand letter is the exception. You generate that one yourself, when you're ready.