About & Legal
This project is a journalistic and public-interest research tool for exploring publicly released government documents.
Purpose & Mission
Epstein File Explorer is an independent, non-commercial research platform built for journalists, researchers, legal professionals, and the public. Its sole purpose is to organize, index, and make searchable the publicly released government documents related to the Jeffrey Epstein investigation.
We do not editorialize, draw conclusions, or assert guilt. We provide tools to navigate what is already public record.
Legal Basis
Every document in this archive originates from official public sources. The legal foundations for this project include:
17 U.S.C. § 105 — Government Works
Works produced by the United States Government are not subject to copyright protection. Documents created by federal agencies (DOJ, FBI, BOP, CBP) and released to the public are in the public domain.
Freedom of Information Act (5 U.S.C. § 552)
FOIA guarantees the public's right to access federal agency records. Documents released under FOIA are public records that may be freely redistributed and analyzed.
Epstein Files Transparency Act (PL 119-38)
Signed into law in 2025, this act specifically mandates the disclosure and public release of federal records related to Jeffrey Epstein. Congress directed the Department of Justice to make these records available to the public.
First Amendment — Right of Public Access
The First Amendment protects the right to access and republish judicial records. Courts have long recognized a constitutional right of access to court proceedings and records, particularly in cases of significant public interest. See Press-Enterprise Co. v. Superior Court, 478 U.S. 1 (1986); Nixon v. Warner Communications, 435 U.S. 589 (1978).
Common Law Right of Access
Under common law, there is a presumptive right of public access to judicial records and documents. Court filings from cases such as Giuffre v. Maxwell (15-cv-07433, S.D.N.Y.) were specifically unsealed for public access by court order in January 2024.
Congressional Records
Documents released by the House Oversight Committee through congressional investigations are public records. Congressional records are protected under the Speech or Debate Clause (Art. I, § 6) and are freely publishable.
Appearance in these documents does not imply wrongdoing, guilt, or involvement in any illegal activity. Many individuals named in these records are witnesses, victims, legal counsel, law enforcement officers, government officials, or persons mentioned in passing. Names extracted by automated systems may include errors or false matches. Always verify information against original source documents.
This platform does not make accusations or allegations. Entity extraction and relationship mapping are automated processes based on co-occurrence in text and should not be interpreted as evidence of any relationship beyond what is documented in the source material.
Methodology
Our data pipeline processes documents through several stages:
- Document Registry — Files are cataloged with SHA-256 checksums for deduplication and integrity verification. Each file is tracked to its original source URL.
- Text Extraction — Text is extracted from PDFs using AI-assisted OCR. Each extraction maintains provenance records linking extracted text to specific page ranges in the source document.
- Entity Extraction — Named entities (people, organizations, locations) are identified using large language models and linked back to their exact position in source text.
- Entity Resolution — Duplicate entities are merged using fuzzy matching with conservative thresholds to prevent false merges.
- Verification — A two-layer verification system scores each entity mention and relationship: first via automated evidence matching, then via adversarial LLM review for uncertain cases. Verification badges are displayed throughout the interface.
- Semantic Search — Vector embeddings enable AI-powered semantic search across the full corpus of extracted text and relationship descriptions.
The source code for the entire pipeline is open source and available for audit. We encourage researchers to verify our methodology and report any issues.
Document Sources
All documents come from the following official public sources:
What This Platform Does NOT Do
- ×Does not host original documents. We index metadata and extracted text. Original files remain with their official sources.
- ×Does not make accusations. Entity connections reflect co-occurrence in documents, not allegations of wrongdoing.
- ×Does not publish private information. Only information already in the public record through official releases is included. We respect government redactions.
- ×Does not provide legal advice. Nothing on this platform constitutes legal counsel or should be relied upon for legal decisions.
- ×Does not guarantee accuracy. Automated text extraction (OCR) and entity recognition are imperfect. Always verify against original source documents.
API & Data Usage Terms
This platform provides a read-only search interface. If API access is made available, the following terms apply:
- •Non-commercial use. The API and search tools are provided for research, journalism, and educational purposes. Commercial use without permission is prohibited.
- •No bulk scraping. Automated mass downloading of data is not permitted. Reasonable programmatic access for research is acceptable.
- •Attribution required. If you publish findings based on data from this platform, please cite the original government source and this tool.
- •No harassment. Data from this platform must not be used to harass, threaten, defame, or doxx any individual. Violations may result in access being revoked.
- •As-is, no warranty. This service is provided "as is" without warranties of any kind. We are not liable for damages arising from use of this platform or decisions made based on information found here.
Open Source
The source code for this platform is released under the PolyForm Noncommercial License 1.0.0. This means:
- • The code can be freely used, modified, and distributed for any noncommercial purpose
- • Permitted uses include personal research, journalism, education, nonprofits, and government institutions
- • Commercial use requires separate authorization from the licensor
- • The underlying documents are U.S. government works in the public domain
- • The entity database and knowledge graph are derived works from public domain sources
- • No data files, API keys, or credentials are included in the repository
Victim Privacy
This project respects the privacy of victims. We voluntarily remove personal identifying information (such as Social Security numbers, financial account numbers, and medical records) that was clearly released in error by government agencies. Government redactions in source documents are always respected.
If you believe your personal information appears on this platform in error, or if you are a victim and wish to request review of specific content, please contact the maintainer for prompt review.
Independent Project
This is an independent journalistic and archival project. It is not affiliated with, endorsed by, or connected to any court, government entity, law enforcement agency, political party, or party to any legal proceeding. Documents are presented as-is for informational and research purposes only.
This site is intended for adults (18+) engaged in research, journalism, or education. The materials contain references to serious crimes including sexual abuse and trafficking.
Responsible Use
These documents exist because Congress, the courts, and federal agencies determined the public has a right to see them. We believe transparency and accountability are essential to a functioning democracy. At the same time, these records describe serious crimes with real victims. We ask all users to approach this material with the gravity it deserves.
If you or someone you know has been affected by trafficking or sexual abuse, the National Human Trafficking Hotline can be reached at 1-888-373-7888 and RAINN's National Sexual Assault Hotline at 1-800-656-4673.
Last updated: February 2026