Epstein-Data

PostedAugust 5, 2026

UpdatedSeptember 6, 2026

ByEpsteinWiki

Epstein Data is a free, independently operated, open source research platform that makes millions of pages from the publicly released Epstein files searchable and easier to examine. As an EpsteinWiki research and evidence partner, Epstein Data provides the document infrastructure that allows contributors and readers to move from an EpsteinWiki claim directly to the underlying EFTA record.

Snapshot

Detail	Information
Name	Epstein Data
Website	epstein-data.com
Project type	Independent public interest research database
EpsteinWiki relationship	Research and evidence partner
Cost	Free public access
Accounts required	No account or login required for ordinary research
Core collection	Approximately 1.4 million documents and 2.9 million pages
Main source	Publicly released United States Department of Justice Epstein files
Document groups	12 Department of Justice datasets, plus additional indexed collections
Search capabilities	Full text, exact phrase, EFTA number, images, entities, transcripts, and structured data
Research tools	Artificial intelligence research assistant, reverse image search, flight index, network graph, geographic map, document viewer, and evidence map
Open source access	Structured data and research repositories are publicly available through GitHub
Government affiliation	None
Required caution	Automated text, entity matches, image descriptions, and analytical reports must be verified against original records

What Is Epstein Data?

Epstein Data is a searchable research platform built to organize the enormous volume of material released in connection with Jeffrey Epstein, Ghislaine Maxwell, related investigations, court proceedings, and government document productions.

The Epstein Data homepage describes the collection as containing approximately 1.4 million documents and 2.9 million pages across 12 Department of Justice datasets. The collection includes scanned documents, photographs, spreadsheets, emails, audio, video, transcripts, investigative records, court materials, and other public records.

The platform converts a difficult collection of separate files into searchable databases. Researchers can search for a name, organization, location, phrase, EFTA identifier, aircraft, financial term, or other research lead.

Epstein Data is independently operated. It is not affiliated with the United States Department of Justice, Federal Bureau of Investigation, any other government agency, or Anthropic.

Collection totals and tool counts may change as records are added, corrected, reprocessed, or classified.

The EpsteinWiki and Epstein Data Partnership

EpsteinWiki recognizes Epstein Data as a research and evidence partner because the two projects serve complementary purposes.

Epstein Data provides large scale document search, direct document access, structured datasets, analytical tools, and technical infrastructure. EpsteinWiki provides researched profiles, timelines, evidence summaries, legal context, editorial review, source explanations, and connections between people, organizations, places, cases, and documents.

The partnership creates a practical research path:

A researcher discovers a document or pattern through Epstein Data.
The original document is opened and reviewed.
The EFTA identifier, page, context, and limitations are recorded.
EpsteinWiki contributors compare the record with court filings, testimony, public records, and other evidence.
The verified information is incorporated into an EpsteinWiki article.
Readers can follow the citation back to the underlying document on Epstein Data.

The relationship is also visible through the Epstein Data Community Resources directory, which identifies EpsteinWiki as a collaborative knowledge base covering people, documents, court cases, and evidence.

Partnership does not mean that every automated result, report, or interpretation is automatically adopted by EpsteinWiki. EpsteinWiki applies its own investigative, privacy, sourcing, and editorial standards before publication.

Why Epstein Data Matters

The public Epstein document collection is too large for conventional page by page review alone. Important references may appear in duplicate records, attachments, handwritten notes, forwarded emails, transcripts, spreadsheets, or poorly scanned pages.

Epstein Data helps researchers:

Locate records across millions of pages
Search exact phrases and name variations
Resolve EFTA identifiers
Compare duplicate documents
Find references across different datasets
Review images and extracted image descriptions
Search audio and video transcripts
Examine flight records
Explore geographic references
Identify possible relationships for further verification
Trace records cited in investigative reports
Download structured data for independent analysis

The platform does not eliminate the need for human research. It makes human review possible at a scale that would otherwise be extremely difficult.

The Document Collection

The Full Text Corpus contains approximately 1.43 million indexed documents and 2.9 million pages.

The source collection includes material from the Department of Justice production released under the Epstein Files Transparency Act. Epstein Data also indexes other relevant public collections, including House Oversight materials and FBI Vault records where available.

The platform’s open source data repository describes its structured exports as covering all 12 Department of Justice datasets, plus House Oversight estate materials and FBI Vault records.

The corpus includes several types of information:

Original document images
Optical character recognition text
Native digital files
Email metadata
Scanned text extracts
Handwriting transcriptions
Audio and video transcripts
Spreadsheet data
Photograph analysis
Document classifications
Redaction analysis
Entity extractions
Related document matches
EFTA statements
Report citations

These information layers serve different purposes. An original document image is evidence. Optical character recognition, an entity extraction, an image description, or an artificial intelligence summary is an aid for locating and interpreting that evidence.

EFTA Document Access

EFTA identifiers are production or Bates numbers assigned to records in the public document release.

A typical identifier appears as:

EFTA00074206

Epstein Data allows direct document access using the identifier:

https://epstein-data.com/EFTA00074206

The EFTA00074206 document page is an example of the direct link format.

A direct EFTA page may provide:

The source document
Page images
Extracted text
Document metadata
Download access
Dataset information
Related records
Searchable content

An EFTA number does not certify that every statement in the document is accurate. It identifies the produced record. Researchers must still determine who created the document, what type of record it is, what it says, whether it is complete, and how much evidentiary weight it deserves.

Required EpsteinWiki Citation Method

Every EpsteinWiki claim based on an EFTA record should link directly to the relevant Epstein Data document.

Use:

EFTA00074206, brief description of the record

Do not use only:

The Epstein Data homepage
A general search page
An entity page
An artificial intelligence response
An extracted database row
An investigation report
A screenshot without the original document
A social media post discussing the record

The citation should include:

Complete EFTA number
Brief document description
Relevant page
PDF page when different
Printed page when available
Document date
Source or author when known
Necessary evidentiary limitation

Place the direct receipt near the claim it supports. Include important direct EFTA links again in the Sources section.

Search Tools

The Epstein Data search system allows researchers to search by name, keyword, exact phrase, or document identifier.

Keyword Search

A keyword search can locate pages containing a term, surname, company, property, address, or subject.

Search common variations, including:

Full name
Surname
Initials
Alternate spelling
Former name
Nickname
Company name
Email address
Telephone number
Street address
Aircraft registration
Trust or corporate entity

A missing search result does not prove that the information is absent. Optical character recognition may fail when documents contain handwriting, poor scans, unusual fonts, damaged pages, or heavy redaction.

Exact Phrase Search

Quotation marks can be used to search for an exact phrase:

“wire transfer”

Exact phrase searching can help locate repeated wording, duplicate correspondence, legal language, addresses, or distinctive quotations.

EFTA Search

Researchers can enter a complete EFTA identifier to locate a specific document.

Always compare the identifier digit by digit. A single incorrect number may lead to a different record or no result.

Database Search

The Explore Databases section allows researchers to search individual collections, including full text, redaction analysis, extracted entities, photographs, audio transcripts, and optical character recognition text.

Results from one database should be compared with results from other relevant collections.

Ask AI Research Assistant

The Epstein Data Research Assistant allows researchers to ask questions in plain language. The assistant searches across the indexed databases and returns a synthesized response with document citations.

The tool can help:

Develop search terms
Locate possible EFTA records
Identify names or entities for review
Compare references across databases
Find relevant document clusters
Generate preliminary research paths
Summarize large result sets
Suggest related records

The assistant is a discovery tool, not an evidentiary source.

Epstein Data states that artificial intelligence summaries may contain errors, hallucinations, or misinterpretations. Every claim must be verified against the linked source document before it is cited or published.

Do not enter personal, private, confidential, or survivor identifying information into the research assistant. Queries are processed through outside artificial intelligence services.

Reverse Image Search

The Reverse Image Search tool allows a researcher to upload an image and search for visually similar material across the indexed image collection.

It can help locate:

Another copy of a photograph
The source document containing an image
A wider version of a cropped photograph
Similar locations or scenes
Repeated evidence photographs
Possible appearances of the same person
Related image sequences

A visual match is a lead. It is not automatic proof that two people, places, or objects are identical.

Researchers should compare:

Facial features
Clothing
Background objects
Architecture
Lighting
Image dimensions
Cropping
Source documents
Dates
Captions
Metadata
Surrounding pages

Do not identify a person solely from an automated face match. Do not use image tools to identify a protected survivor, witness, or minor.

Flight Network

The Epstein Data Flight Index maps thousands of flight records and allows research by passenger, route, aircraft, date, and destination.

The index can help researchers:

Find passenger name variations
Reconstruct flight sequences
Compare routes
Identify aircraft
Review departure and arrival locations
Examine repeated travel patterns
Locate the original source records

A database entry remains a transcription or structured interpretation of a source record. Researchers must inspect the underlying log or manifest before publication.

A flight entry may document that a person was listed on a particular leg. It does not establish why the person traveled, what happened at the destination, or whether the passenger knew about another person’s conduct.

Network and Relationship Tools

The Epstein Data network system organizes people, organizations, locations, and documented relationships into searchable visual structures.

The tools can help identify:

Repeated correspondence
Shared corporate roles
Travel overlap
Financial references
Legal relationships
Employment
Scheduled meetings
Document co appearances
Property connections
Possible research clusters

A network graph is a research map. It is not a guilt chart.

Two people may appear near each other because their names occur in the same document, because they shared an institution, or because an automated system proposed a relationship. Every connection requires review of the underlying evidence.

EpsteinWiki should describe the exact relationship instead of relying on vague language such as “connected to Epstein.”

Geographic Research

The Global Heatmap organizes geographic references across the corpus. The platform reports references involving 145 countries.

Geographic tools can assist with:

Location searches
Property research
Travel patterns
International corporate records
Government contacts
Institutional networks
Event reconstruction
Regional document clusters

A country appearing in a document does not establish government involvement. A location count may include travel references, addresses, news articles, legal citations, business records, or incidental mentions.

Researchers must open the cited documents and determine what each location reference actually represents.

FBI Evidence Map

The FBI Evidence Map reconstructs serial numbers, evidence receipts, witness material, and other records associated with FBI investigations.

The map can help researchers compare:

Indexed records
Located documents
Missing documents
Illegible references
Evidence receipt numbers
FBI serial numbers
Testifying witness material
Non testifying witness material
Related case numbers

A record marked as missing means the project did not locate the expected item within the public release under the stated method. It does not necessarily prove that the government destroyed, concealed, or never possessed the record.

Missing record analysis should clearly identify the index, expected serial, search method, and limits of the public production.

Investigation Reports

The Epstein Data Investigation Reports library contains forensic reports organized across multiple research categories.

Topics include:

Financial forensics
Institutional failures
Congressional briefings
Evidence and device analysis
Intelligence records
Social network analysis
Government officials
Dataset analysis
Internet theories
Prosecutorial questions
Document removal
Missing records
Properties and travel
Legal records
Redaction issues

Each report is intended to connect analytical findings with specific EFTA citations.

Epstein Data states that its analytical text and investigation reports are generated with artificial intelligence and iteratively checked against source documents, but may still contain errors. EpsteinWiki contributors must independently verify any report before using it.

The correct research sequence is:

Read the report.
Identify the relevant claim.
Open every cited EFTA record.
Verify the quotation and page.
Review surrounding pages.
Search for contradictory evidence.
Check related court or government records.
Write an independent EpsteinWiki assessment.

Do not copy an automated report into EpsteinWiki and present it as independently verified research.

Survivor Voices

The Hear From the Survivors section collects publicly available video interviews and statements from survivors who spoke publicly and on the record.

The page centers survivor accounts as the reason the records matter. It links to recordings hosted by their original publishers rather than reproducing complete testimony as extracted text.

EpsteinWiki contributors using this section must:

Respect the survivor’s public framing
Preserve the context of the interview
Verify quotations against the recording
Include timestamps
Avoid unnecessary graphic detail
Avoid inferring consent
Avoid identifying protected people
Link to the original publisher when available
Follow current survivor privacy requests

A survivor’s public interview is an important primary account. It should be treated with dignity and evaluated according to its context, subject matter, and relationship to other evidence.

Open Source Data and Independent Review

Epstein Data publishes structured exports through the Epstein Research Data GitHub repository.

Available resources include structured information derived from:

Full text documents
EFTA mappings
Entity extractions
Knowledge graphs
Images
Redaction analysis
Communications
Transcripts
Spreadsheets
Document classifications
Metadata
Investigation report citations

Open source access allows researchers to inspect the data structure, reproduce searches, perform independent analysis, and build additional public interest research tools.

Structured data can contain processing errors. Researchers should treat database rows as pointers to evidence and verify them against the source file.

Tools for Developers and Advanced Researchers

Epstein Data provides technical access for researchers who need to work with large collections.

Available options include:

Public structured data
Downloadable databases
A REST API
OpenAPI documentation
Model Context Protocol access
Local database installation
Command line research tools
Artificial intelligence agent instructions

The Run Your Own Investigator guide explains how researchers can download the databases and search them locally.

Local analysis may provide greater control over search methods and research notes. It does not remove the obligation to protect private information or verify original records.

Privacy Practices

The Epstein Data Privacy and Terms page describes the site as a public interest research tool.

According to the published policy:

Ordinary use does not require an account
The application does not use Google Analytics
Search queries are recorded without an application level IP address or browser fingerprint
Recent artificial intelligence chat history is stored in the user’s browser unless the user chooses to share it
Artificial intelligence questions are processed through an outside service
The platform provides a privacy and removal request process
Users are instructed not to republish information that may identify a trafficking or abuse victim

Infrastructure providers may maintain standard security logs. Researchers should read the current privacy policy before submitting sensitive material or using interactive tools.

The Epstein Data contact page includes options for general feedback, privacy concerns, removal requests, document leads, broken links, and press inquiries.

Privacy and Takedown Requests

Epstein Data provides a specific process for reporting:

Mistaken redactions
Exposed personal information
Survivor identifying information
Photographs of minors
Nonconsensual intimate imagery
Other privacy concerns

A report may identify the EFTA number, page, URL, and nature of the concern.

EpsteinWiki contributors who locate exposed survivor or minor information should not repost it. They should preserve only what is necessary to document the issue and report it through the appropriate privacy channel.

See the EpsteinWiki Privacy Safeguards for Minors before handling sensitive material.

What Epstein Data Can Establish

Epstein Data can help establish that:

A particular document appears in an indexed public collection
A term appears in extracted or searchable text
A document is associated with an EFTA identifier
A record contains a particular image or statement
Multiple copies or versions of a document may exist
A name or entity appears within a database
A source record may support further investigation

The original document, not the search result, establishes the underlying evidence.

What Epstein Data Does Not Automatically Establish

A search result does not automatically prove:

A person’s identity
A personal relationship
Attendance at a scheduled meeting
The purpose of travel
Knowledge of criminal conduct
Participation in criminal conduct
Accuracy of a witness statement
Authenticity of every attachment
Beneficial ownership of an entity
The meaning of a redaction
That a missing record was intentionally withheld
That an artificial intelligence summary is correct
That an extracted entity match refers to the intended person

Names in the Epstein files can belong to survivors, witnesses, investigators, attorneys, employees, social contacts, public officials, journalists, medical professionals, unrelated third parties, or people mentioned only in news material.

Document presence must never be converted into guilt by association.

Known Research Limitations

Researchers should account for several limitations.

Optical Character Recognition Errors

Scanned pages may produce incorrect words, names, dates, or numbers. Search the original image and common spelling variations.

Handwriting Errors

Automated handwriting transcription may misread initials, names, telephone numbers, or short notes.

Entity Matching Errors

Two people with similar names may be merged. One person may also appear under several spellings.

Image Description Errors

Automated image descriptions may misidentify people, places, objects, or activities.

Artificial Intelligence Errors

Generated summaries may hallucinate facts, combine separate records, omit qualifications, or misunderstand legal language.

Incomplete Public Releases

A search covers the indexed material, not every record that may ever have existed. Absence from the database is not proof that an event did not occur.

Duplicate Records

The same email, filing, image, or report may appear repeatedly. Duplicate appearances do not create independent corroboration.

Redactions

Redacted information should not be guessed. Optical character recognition near a redaction may produce meaningless text rather than recovered content.

EpsteinWiki Verification Workflow

Step 1: Search Broadly

Search the name, phrase, entity, address, or EFTA identifier across relevant Epstein Data collections.

Step 2: Open the Original Record

Do not rely on a result preview, extracted row, or artificial intelligence summary.

Step 3: Verify the Identifier

Confirm the complete EFTA number and document page.

Step 4: Read the Entire Context

Review the full page, surrounding pages, attachments, email thread, or related document sequence.

Step 5: Classify the Record

Determine whether the material is a court filing, witness statement, investigative note, contact entry, email, calendar entry, photograph, financial record, news clipping, or another record type.

Step 6: Define What It Establishes

Write the narrowest accurate statement supported by the document.

Step 7: Identify Limitations

State what the record does not prove.

Step 8: Seek Corroboration

Compare the record with other EFTA documents, court filings, public records, testimony, and reliable reporting.

Step 9: Search for Contradictions

Look for denials, conflicting dates, alternate identities, later filings, corrections, and reasonable alternative explanations.

Step 10: Protect Survivors and Minors

Check names, images, metadata, indirect identifiers, and redactions before saving or publishing material.

Step 11: Add the Direct Citation

Link the complete EFTA identifier directly to Epstein Data near the relevant claim.

Step 12: Conduct Editorial Review

Verify the article against the original evidence before publication.

Questions

Is Epstein Data an official government database?

No. It is an independent public interest research project. The underlying collection includes government records, but the platform is not operated by the Department of Justice, FBI, or another government agency.

Is Epstein Data free?

Yes. The public research interface is available without a subscription or required account.

Why does EpsteinWiki use Epstein Data?

Epstein Data provides searchable access and stable direct links to millions of pages. This allows EpsteinWiki to connect documented claims with the records supporting them.

Does the partnership mean that every Epstein Data report is approved by EpsteinWiki?

No. EpsteinWiki independently verifies records and applies its own investigative and editorial standards.

Can an Epstein Data artificial intelligence answer be cited as evidence?

No. Use the answer to locate relevant documents. Cite and analyze the underlying records.

Is an EFTA number proof that a statement is true?

No. It identifies a record in the production. Statements inside the record must still be evaluated according to their source, context, and evidentiary status.

Can a search result prove that two people had a relationship?

No. The underlying documents must establish the nature, timing, and significance of the connection.

What should a researcher do when optical character recognition conflicts with the document image?

Use the document image. Record the transcription error when it could affect searches or interpretation.

Does no search result mean that no evidence exists?

No. A name may be misspelled, handwritten, redacted, poorly scanned, stored in another collection, or absent from the public release.

Can a reverse image result identify someone?

It can generate a lead. Identity must be independently verified before publication.

Are Epstein Data investigation reports human written?

The platform states that its analytical reports and text use artificial intelligence and are iteratively checked against source documents. Errors may remain, so EpsteinWiki requires independent verification.

Can researchers download the data?

Yes. Epstein Data provides open source repositories, structured exports, technical documentation, and options for local research.

What should a contributor do after finding an important EFTA document?

Open the original, verify the identifier and pages, read the surrounding context, preserve the direct link, seek corroboration, assess privacy, and submit the finding through the EpsteinWiki editorial process.

What should a contributor do after finding exposed survivor information?

Do not republish it. Record only what is required to locate the problem and submit a privacy report through Epstein Data and the appropriate EpsteinWiki safety channel.

Can Epstein Data replace court record research?

No. Court dockets should be checked directly for filing history, later orders, amendments, appeals, dismissals, and final outcomes.

Who operates Epstein Data?

The platform describes itself as an independently run research project. Its public source data and research repositories are maintained through the rhowardstone GitHub account, and contact with the operator is available through the site’s contact page.

Related EpsteinWiki Guides

Sources

Editorial Note

Epstein Data is an EpsteinWiki research and evidence partner. EpsteinWiki independently reviews all records, analytical claims, automated outputs, and investigation reports before incorporating them into published articles. Collection totals and available tools may change as Epstein Data processes additional records and improves its databases.

Last reviewed: September 6, 2026

Searchable Epstein Databases