Epstein Files: AI-Powered Document Search

A new public transparency project has launched at epstein-files.org, offering free semantic search across tens of thousands of documents related to the Jeffrey Epstein case—complete with AI-generated podcasts that explain complex topics.

Built by Dr. Andrew Walsh, MD/PhD, the site provides professional-grade document intelligence tools typically reserved for enterprise applications. The system uses AI-powered semantic search and entity extraction to make the massive document corpus navigable—something that would be nearly impossible through manual review.

Key features include:

• 33,891+ documents processed with AI document intelligence • Semantic search that understands context, not just keywords • AI-generated podcasts that summarize popular searches and complex topics • AI summaries with source citations for quick research • Export functionality to analyze results with ChatGPT or Claude • Free public access with no paywalls or registration

The podcast feature is particularly innovative—automatically generating audio explanations of frequently asked questions and research topics, making dense legal documents accessible to a broader audience.

The creator emphasizes an important research disclaimer: AI summaries and podcasts should be used for initial research only. Users are encouraged to verify findings by searching original source documents using the keyword search functionality.

This represents over 200 hours of unpaid development work, offered as both a public service and a demonstration of professional document intelligence system architecture. The site accepts donations to help cover hosting costs and continued maintenance.

For researchers, journalists, and anyone interested in transparency, epstein-files.org provides a powerful tool for navigating this significant public record.

Medium link: https://medium.com/@tsardoz/i-made-33-891-sealed-epstein-doc...

33,891. That’s a lot. Are there more?