EDA

Ask the Epstein Files

Ask questions about 207,251+ documents from the Jeffrey Epstein investigation. The AI searches relevant documents and provides cited answers with direct links to source materials, page numbers, and PDF files. Powered by retrieval-augmented generation for accuracy.

Ask the Epstein Files

Ask questions about the 207,000+ documents in the archive. The AI will search relevant documents and provide cited answers.

AI Q&A Coming Soon

AI Q&A will be available once the Anthropic API is configured. You can still type questions to see the interface in action.

Suggested Questions

AI responses are generated from the archive documents. Always verify with primary sources.

How It Works

1

Ask a Question

Type any question about the Epstein investigation, documents, people, or events.

2

AI Searches Documents

The AI searches across all 207,251 documents using semantic search to find relevant materials.

3

Get Cited Answers

Receive an answer with citations linking to the specific documents and page numbers used.

Frequently Asked Questions

How does the AI Q&A work?
The AI Q&A system uses retrieval-augmented generation (RAG) to search across 207,251+ documents in the archive. When you ask a question, the system finds the most relevant documents using semantic search, then generates an answer based on those documents. Every response includes citations with links to the source documents.
Are the AI answers accurate?
The AI generates answers based strictly on the documents in the archive. It does not fabricate information or use external knowledge. However, AI can sometimes misinterpret context or nuance. Every answer includes source citations so you can verify the information against the primary documents.
What types of questions can I ask?
You can ask about people, events, dates, document contents, connections between entities, flight log details, email correspondence, and more. The system works best with specific, focused questions rather than overly broad queries.
Does the AI have access to all documents?
The AI has access to the full text of all 207,251 documents in the archive, including court records, FBI files, depositions, emails, and other materials. However, some documents with poor OCR quality may have incomplete text extraction.
Is my conversation private?
Questions are sent to the AI for processing but are not stored permanently or associated with your identity. The system does not require an account or login. Search queries may be logged anonymously for improving search quality.
What AI model powers this feature?
The Q&A feature is powered by Anthropic's Claude AI model, which is specifically designed for safe, accurate, and helpful interactions. The model is instructed to only answer based on the archive documents and to cite its sources.