Methodology

This page describes how documents are sourced, verified, cited, and presented on this site. It is the principal reference page for the editorial standards we follow. We update it as practice develops.

Sourcing

The principal documentary basis of this site is the U.S. Department of Justice's release of materials under the Epstein Files Transparency Act, made publicly available beginning 30 January 2026 at justice.gov/epstein. The release is organised into twelve Data Sets and contains, on its face, more than 33,000 pages of correspondence and scheduling documents from Jeffrey Epstein's estate.

Where documents bearing on the subjects of our coverage are available through the DOJ portal, we cite them by EFTA ID, link to the canonical justice.gov/epstein URL, and reproduce the document scan on the Sources index page where copyright and DOJ release terms permit.

Where documents bearing on our coverage are not yet available in the DOJ portal but are documented in scanned form by independent third-party aggregators (notably the X user William Barnes, @bahnhofswartz, whose extensive aggregation of the 30 January 2026 release across multiple posts has surfaced documents not returned by our own DOJ portal searches), we cite the third-party source and the document content. Where such documents subsequently appear in the DOJ portal, we update the citation to include the canonical reference and reverify the document text against the canonical source.

Where documents are accessible only through press reporting on undisclosed primary sources (notably the San Francisco Standard's reporting on the 2017 Vespers funding), we cite the press reporting and indicate the limit of the public record.

We do not pay for documents. We do not solicit anonymous source material. We do not work with confidential sources who require source protection. The work on this site is built on documents already in the public record, presented with greater context and analysis than has been the practice of the larger press outlets covering the same material.

Verification

For each document cited on this site, we verify:

  1. Authenticity — that the document appears in the DOJ release or is otherwise from a documented release source. We do not rely on documents that cannot be traced to a documented release.
  2. Date — that the date on the document is consistent with the surrounding correspondence chain and with publicly available information about the named parties' movements at the relevant time.
  3. Sender and recipient — that the named senders and recipients are identifiable and that their identities are consistent with publicly available information.
  4. Subject — that the subject of the correspondence is consistent with the surrounding text and with the public record.
  5. OCR quality — that the scanned text accurately reflects the document's content. Where OCR has produced obvious garbling, we identify the garbling and rely on visual inspection of the scan.

Where verification fails on any of these dimensions, we do not cite the document. Where verification is partial — for instance, where a date is approximate or a sender is partially redacted — we identify the partial verification and limit our claims accordingly.

Citation

Every factual claim in our articles is cited to its source. Citations take three forms:

  1. DOJ documents: cited by EFTA ID with date, parties, and subject. The full text is available on the Sources index page. The canonical link is justice.gov/epstein/files/DataSet%20{N}/{EFTA-ID}.pdf, where {N} is the relevant DataSet number.
  2. Press reporting: cited by publication, date, and reporter where known. We link to the original article. We do not paywall-strip; if the original is paywalled, we note this.
  3. Public statements: cited by speaker, date, and venue. Where the public statement is on a public platform (such as X or Medium), we link to the original. Where it has been removed from the public platform, we note this and link to an archived version where one exists.

Quoting

We follow strict quoting limits in our editorial practice. Direct quotes from any document or public statement are typically kept under twenty words, with most under fifteen. The longer ceiling exists to allow direct reproduction of editorially load-bearing single-sentence statements; the spirit of the rule is paraphrase by default, with quotation reserved for cases where the original phrasing carries weight that paraphrase would lose. We rely on at most one direct quote per source per article; further reference to the source is by paraphrase. Where a paraphrase risks misrepresenting the source's emphasis, we flag this in context.

This is partly an editorial choice (paraphrase compels precision) and partly a copyright caution (extensive reproduction of source material exceeds the bounds of fair use even for journalistic purposes). It does not affect the substance of our coverage.

Naming and right-of-reply

Where we name an individual or institution in an article, we send them — wherever a contact pathway is available — a copy of the relevant article, an identification of the specific claims made about them, and an invitation to respond on the record. We allow seven days for a response from the date of sending.

Responses received are integrated into the article body, footer, or Right of reply page as the substance warrants. Where a named party declines to respond, this is noted in the article. Where no response is received within the window, this is noted. Where a response is received after the window, we update the article and note the late response.

We do not name individuals whose involvement in the documentary record is incidental and whose identification adds nothing to the editorial substance. We do not publish home addresses, family details, or other personal information beyond what is in the documentary record itself.

What we don't do

We do not allege criminal conduct that the documents do not establish. We do not impute states of mind. We do not claim that named individuals knew about specific Epstein crimes at specific times unless the documents themselves establish this. We do not call for boycotts, removals, or de-platformings.

We do not opine on the artistic, scientific, or commercial merit of any named individual's work, except in narrow context where the merit is itself a documentary fact (for instance, that a work is in a museum's permanent collection).

We do not engage in personal commentary about the named parties, their families, or their associates beyond what the documentary record supports.

We do not run advertisements or sponsored content. We do not accept paid promotion. The site has no commercial revenue model and is funded out of editorial pocket. The decision not to monetise is intentional and is a load-bearing element of our editorial independence.

Errors and corrections

We make errors. When we discover them, we correct them transparently. The Corrections page lists every correction we have made, with date, the original error, and the corrected text. We do not silently revise published articles to obscure prior errors.

If you have identified an error in our coverage, please contact us via the contact pathway at the foot of every page. We respond to all error reports within seven days.

Editorial independence

This site is operated by an independent editorial team. We have no financial, professional, or personal relationship with any party named in our coverage. We do not accept gifts, hospitality, or paid travel from any source. We do not maintain off-the-record sourcing arrangements with named parties or their representatives.

The named parties have no editorial control over this site, no advance access to our drafts (beyond what the right-of-reply process provides), and no veto over our coverage. We do not negotiate the terms of our coverage in advance with named parties.

Editorial team

The editorial team operating this site is currently working pseudonymously, on advice and for the safety of contributors. We acknowledge that pseudonymous journalism carries reputational costs. We accept those costs in light of the practical risks of named journalism on this specific subject. The decision to work pseudonymously is reviewed regularly and may be reversed.

The principal editor of this site has more than fifteen years of experience in long-form investigative writing, has been published in major U.S. and U.K. outlets, and is supported by a small editorial team with documented experience in legal review, archival research, and design. None of the team has any conflict of interest in respect of any party named in this site's coverage.

Updating this page

We update this methodology page as our practice develops. Significant changes will be flagged on the page itself with a "last updated" stamp.

Last updated: 29 April 2026.