Epstein Files

Searchable text versions of all PDFs released under the Epstein Files Transparency Act (H.R. 4405). The DOJ's OCR is inaccurate and difficult to bulk download, so we reprocessed every file using Tesseract OCR.

Transcripts of audio and video files and higher-accuracy OCR results from DeepSeek-OCR-2 will be made available on this site soon.

Download OCR Files

These files have been processed with optical-character-recognition (OCR) using Tesseract. Click on any file to download it.

Data Set 1
3,112 documents•3,112 pages•0.1 MB
Data Set 2
576 documents•700 pages•0.1 MB
Data Set 3
68 documents•1,847 pages•0.3 MB
Data Set 4
153 documents•2,705 pages•3.0 MB
Data Set 5
121 documents•121 pages•0.1 MB
Data Set 6
14 documents•487 pages•0.4 MB
Data Set 7
18 documents•661 pages•0.7 MB
Data Set 8
11,031 documents•27,302 pages•35.4 MB
Data Set 9Still processing
104,719 documents•336,395 pages•412.1 MB
Data Set 10
503,099 documents•898,310 pages•825.9 MB
Data Set 11
331,640 documents•515,310 pages•467.6 MB
Data Set 12
153 documents•1,526 pages•1.5 MB

Want More Powerful Search?

Standard Works is building next-generation eDiscovery tools with advanced AI-powered search, analysis, and document review. Our eDiscovery platform will enter public beta on or before February 14th. If you would like to get early access, join our waitlist. Priority will be given to journalists, lawyers, and entities representing victims.

These documents are public records. Standard Works provides this service to facilitate research and journalism.

If you are a victim and believe your personal information appears in these documents, we want to help. We can use our eDiscovery tools to help locate your PII, redact it from our files, and assist you in requesting redaction from the DOJ. Please contact william@standardworks.ai.

Disclaimer: The text files provided are outputs from ML models. We make no legal claims regarding the accuracy or completeness of these transcriptions. The majority of documents in this collection were sourced directly from the Department of Justice, while others were obtained through trusted crowdsourcing efforts. For authoritative text, please refer to the DOJ's official website.