How ScrubMyCase works
No magic, no black box. Here's exactly what happens to your document — and what happens to it afterward (it gets deleted).
- 1
You upload & answer three questions
Document type, your role, and your state. That context tailors everything that follows — a police report is scrubbed differently than a custody filing.
- 2
We read the file
Text-based PDFs are read directly; photos and scanned pages are run through OCR so handwriting-era paperwork still works.
- 3
The Classifier maps the claims
The document is broken into individual statements and events, and a timeline is built.
- 4
The Contradiction Hunter cross-checks everything
Every claim is compared against every other claim, and against what you told us, across all ten lenses.
- 5
The Law Mapper proposes concepts — by name only
It names the legal doctrines your document touches on (like "equitable distribution" or "unconscionability"). It is explicitly forbidden from writing any statute number or case citation.
- 6
The Fact-Check agent grounds every one against real sources
This is the anti-hallucination step. Each concept is matched against a hand-verified legal knowledge base and corroborated against real, public legal sources — the official US Code, the Code of Federal Regulations (eCFR), and the CourtListener case-law database. Every item we show links to a source you can open and read. Anything we can't confirm is shown as a "general concept" with a link to search the real source — never an invented citation. A final filter strips any stray fake citation from the text.
- 7
The Report Writer puts it in plain English
You get a calm, color-coded report — every finding, with the quote and where it is, plus the laws that may apply.
🔒 And then your document is deleted.
We don't keep your file, we don't train AI on it, and we don't sell your data. The report link stays live for 24 hours so you can save it.