This repository contains markdown conversions of government-released documents related to Jeffrey Epstein from various sources.
All documents have been processed through Mistral OCR, which is (in my testing) currently the best document-aware processing system available. The OCR is context-aware and can recognize redactions and other document features, providing a more easily parsable unstructured dataset than the traditional OCR that shipped with the government's released documents.
- U.S. Department of Justice document library
- FBI document library
- Text only: Images are not preserved in this conversion process
- Filenames unchanged: Original filenames from the source releases are maintained
- Last updated: December 23, 2025
These markdown files provide a cleaner, more accessible format for researchers and analysts working with this document set.