The receipts behind Epstein’s websites, Google manipulation, Wikipedia cleanup, contractor network, and manufactured online identity
Jeffrey Epstein did not simply hire someone to make a nicer website. After becoming a registered sex offender, he funded an online system designed to change what people saw when they searched his name.
The network included exact name domains, science and philanthropy sites, hosted blogs, social accounts, press releases, business directories, Wikipedia activity, favorable articles, and pages about unrelated people who also happened to be named Jeffrey Epstein. The goal was not subtle. Epstein wanted flattering material to compete with reporting about his crimes and the survivors who exposed him.
Part One of Butterfly Bureau Episode 7 follows the website machinery. The evidence comes from contractor communications, payment records, archived pages, search screenshots, sex offender registry filings, domain histories, and Epstein’s own internet account disclosures.
This was not ordinary image polishing. It was an attempt to purchase a replacement version of reality, one search result at a time.
Why the Website Operation Mattered
Search results shape credibility. When someone warns that a powerful man is dangerous, one of the first things people do is search his name. If that search produces university connections, science philanthropy, charity claims, and flattering biographies, the warning can appear less credible.
That makes reputation laundering more than a vanity project. It can become part of the infrastructure that protects an abuser. Epstein’s online operation did not erase the truth. It created enough favorable material to surround it, dilute it, and make the public work harder to find it.
The 2010 Search Suppression Proposal
In September 2010, EFTA02421209 records a discussion between Epstein and Al Seckel about displacing unfavorable search results. The proposed cost was approximately $20,000 to $25,000.
Epstein asked whether positive articles would be needed to push down the negative material. Seckel confirmed that favorable content was part of the plan.
That exchange establishes purpose. The project was not limited to correcting an address or updating an old biography. Epstein wanted negative results moved and understood that the operation required new positive material.
Real Sites, Pseudo Sites, and a Manufactured Search Page
By October 2010, the idea had become a written strategy. EFTA00754635, EFTA00754636, and EFTA00754637 describe two websites that would openly promote Epstein’s claimed support for science and philanthropy.
The plan also proposed three “pseudo sites” built around sports, cooking, and photo or video sharing. Those subjects had little to do with Epstein’s actual public identity. Their value was that they could create additional results containing the name Jeffrey Epstein.
The sports site was designed to appear connected to another person with the same name. The proposed cooking and sharing sites have not been shown to have launched. They remain proposals, not confirmed websites.
The distinction matters. A plan proves intent. It does not automatically prove that every proposed site was built or that every person named in the proposal performed the assigned work.
The People Named in the Website Plan
Al Seckel was presented as the person overseeing the project and communicating with Epstein. Later correspondence describes search work, Wikipedia changes, content promotion, and efforts to reduce the visibility of negative material.
Michael Keesling was assigned responsibility for content connected to the secondary sites. Reporting based on the released correspondence says Keesling hired writers in the Philippines and agreed to perform the work for $25,000.
Pablos Holman, Isabel Maxwell, Christine Maxwell, Paul Kirkaas, and Baron Reichart appeared as possible search, technical, or consulting resources. Their inclusion proves that Seckel presented them as part of a potential team. It does not establish that each person accepted the assignment, performed work, controlled a site, or understood every purpose of the campaign.
Gregory Markel of Infuse Creative was also approached. Reporting based on the released emails says he declined after raising concerns about Epstein’s conviction and the allegations against him. His refusal matters because it shows that contractors could recognize the ethical problem and walk away.
Others made a different choice.
Epstein Said the Money Was Sent
In EFTA00778138, Epstein responded that money had been sent. The record supports the conclusion that he funded the project.
The public files do not yet provide a complete payment chain identifying every recipient, account, invoice, and task. The records prove a funded reputation project, but they do not support inventing a complete ledger where one has not been released.
The Sites Went Live
Within days, the operation reportedly launched JeffreyEpsteinScience.com, JeffreyEpstein.org, and JeffreyEpsteinSports.com.
The science and philanthropy properties presented Epstein as a patron of important ideas and respected researchers. The sports site published material under the name Jeffrey Epstein while reportedly using outsourced writing. Repeated use of his full name helped the pages compete in exact name searches.
The trick was volume. Several pages could look like separate sources while tracing back to one coordinated campaign. A promotional claim could begin on an Epstein site, reappear in a press release, move through a hosted blog or directory, and return to Google looking like independent confirmation.
It was the same press release wearing six different wigs.
The Team Captured Its Google Results
On November 5, 2010, a Google screenshot preserved in EFTA00752512 showed three targeted favorable properties appearing on the first page.
One screenshot cannot establish what every user saw. Search results change by location, time, device, and browsing history. It does prove that the operation monitored Google and captured at least one search in which several targeted sites obtained prominent placement.
That is evidence of a measured campaign, not random web publishing.
Wikipedia Became a Battleground
Wikipedia mattered because its pages ranked highly and appeared authoritative. Correspondence reported by The Guardian describes Seckel celebrating changes to Epstein’s page, including replacing the mugshot, reducing the prominence of criminal information, and expanding material about philanthropy and science.
The correspondence shows that Epstein’s team wanted the page changed and discussed efforts to influence it. It does not prove that every historical edit came from the same operator. Assigning individual edits requires matching accounts, timing, edit histories, and technical evidence.
The goal is documented. Make the criminal record quieter. Make the philanthropy louder. Then hope the respectable fonts do the rest.
The Sex Offender Registry Preserved the Network
One of the most revealing twists is that Epstein’s sex offender registration obligations helped preserve the infrastructure built around him.
EFTA01110814 and EFTA00686364 document the disclosure of the Twitter identity jeffreyepstein1. Epstein later signed a confirmation that he had opened the account.
A January 2012 filing beginning with EFTA00586845 and EFTA00586846 disclosed email addresses, websites, blogs, social profiles, and online accounts.
The larger inventories in EFTA01136491 through EFTA01136494 identify ten Epstein branded websites and six additional properties maintained by third party reputation providers. They also list accounts across Blogspot, Facebook, LinkedIn, Myspace, Pinterest, PR.com, Twitter, Vimeo, Wikipedia, Yahoo, Yelp, YouTube, and numerous directories.
The inventory does not prove that Epstein personally created every account. It establishes a wide distribution footprint associated with the online system reported to registration authorities.
The registry requirement Epstein likely considered an inconvenience became one of the best maps of the machine built to obscure why he was on the registry.
Outside Reputation Providers Created the Infrastructure
In EFTA00306059 and EFTA00306060, attorney Darren Indyke told the Florida Department of Law Enforcement that jeffrey@jeffreyepstein.org had been established by a third party internet reputation provider.
The letter stated that providers independently created internet accounts, websites, blogs, email addresses, and related content concerning Epstein’s scientific and philanthropic activities.
That description is unusually direct. The contractors were not maintaining one official biography. They were creating an ecosystem around a convicted sex offender’s preferred identity.
Reputation Changer Accepted the Campaign
In 2012, Reputation Changer prepared a plan seeking to fill the first four pages of Google, Bing, and Yahoo with favorable material. It proposed websites, articles, press releases, social profiles, directories, backlinks, search optimization, and monitoring.
The correspondence stated that Epstein’s name had been reviewed by the company’s ethics committee before it accepted the work. Apparently the ethics committee located the word “ethics,” admired it briefly, and returned to the invoice.
On October 16, 2012, EFTA01580804 records instructions to wire $12,500 from Epstein’s JPMorgan account to Reputation Changer LLC for a one month trial. The next day, EFTA01993247 provided access instructions for a campaign dashboard.
The proposal, wire instruction, and dashboard email establish that the effort progressed beyond a sales pitch.
Tyler Shears and the Later Search Campaign
Later records connect Tyler Shears to website and search reputation work. Forbes reporting says Shears created jeffreyepstein.net, managed jeffreyepstein.org, and received at least $50,000 for reputation related services.
In 2014, Christina Galbraith instructed a contractor named Tyler to promote exact name pages that were “not our websites.” EFTA01928953 and EFTA01928954 concern pages about unrelated people named Jeffrey Epstein.
The strategy did not require inventing those people. It could use their legitimate pages to make searches for Jeffrey Epstein return results about someone else.
If the truth could not be removed, make the name itself noisy.
The Content Factory
EpsteinWiki’s investigation of the website network identifies eighteen historically supported domains and hosted sites. The operation included foundation pages, science sites, education and finance pages, WordPress and Blogspot accounts, press release services, social profiles, business directories, syndicated articles, and Wikipedia efforts.
These were not automatically independent sources. Many pages cross linked, repeated the same claims, shared addresses, used related technology, or circulated material that originated inside Epstein’s network.
Repetition is not corroboration. Ten search results can still be one man’s public relations department yelling into ten different microphones.
What the Receipts Establish
The evidence establishes that Epstein financed efforts to displace negative search results with favorable content. Seckel discussed and oversaw a 2010 campaign. A written plan proposed genuine promotional sites and deceptive pseudo sites. Keesling was assigned a content role and reportedly hired outside writers. Several other people were proposed as technical or search resources. At least three targeted sites went live, and the operation monitored their Google placement.
The records also establish that Epstein used outside reputation providers, paid Reputation Changer for a trial campaign, maintained a wide network of domains and distribution accounts, and later used contractors including Tyler Shears in website and search promotion.
What the Receipts Do Not Establish
The records do not prove that every person named in a proposal accepted the work. They do not identify the creator of every account, the author of every post, or the person behind every Wikipedia edit. They do not establish that current owners of expired domains are connected to Epstein. They do not provide a complete financial ledger.
The website evidence also does not establish that every contractor participated in Epstein’s trafficking or abuse. The documented conduct is reputation work performed for, proposed to, or discussed with a convicted sex offender. That conduct is serious enough without stapling unsupported crimes to it.
Follow the Receipts
Epstein’s online reputation was not cleaned by one clever hacker pressing delete. It was rebuilt through websites, contractors, writers, search tactics, Wikipedia pressure, social accounts, directory listings, payments, and institutional credibility.
Survivors were telling the truth about what he did. Epstein was paying to change what the public saw when it looked for confirmation.
He spent money trying to control what the internet remembered. The receipts make sure the internet remembers the evidence.
Primary Evidence
- EFTA02421209 records the September 2010 search suppression discussion.
- EFTA00754635 through EFTA00754637 preserve the October 2010 website strategy.
- EFTA00778138 records Epstein saying the money had been sent.
- EFTA00752512 preserves the November 2010 Google results screenshot.
- EFTA00586845 and EFTA00586846 begin the January 2012 internet inventory.
- EFTA01136491 through EFTA01136494 document the broader website, account, and contractor inventory.
- EFTA00306059 and EFTA00306060 contain Darren Indyke’s description of outside reputation providers.
- EFTA01580804 records the $12,500 Reputation Changer wire instruction.
- EFTA01993247 records access to the Reputation Changer campaign dashboard.
- EFTA01928953 and EFTA01928954 concern promotion of unrelated people named Jeffrey Epstein.
Supporting Sources
- EpsteinWiki’s complete website network investigation documents the domains, archived pages, contractors, current owners, and evidence limits.
- The Verge’s investigation examines the 2010 search strategy and online cleanup campaign.
- Forbes reporting on Epstein’s search operation documents Tyler Shears and later website work.
- The Guardian’s reputation campaign report documents the Wikipedia and search cleanup correspondence.