The Power of Cross-Indexing: How 37 Indexes Exposed What Single-Source Searches Can't
- Patrick Duggan
- Feb 27
- 7 min read
# The Power of Cross-Indexing: How 37 Indexes Exposed What Single-Source Searches Can't
**One name. Thirty-seven indexes. 10.9 million documents. The connections only appear when you search everything at once.**
Today we ran a deep scan on Donald Trump across every index in our system. We searched every known alias. We pulled every co-occurrence. We ran five intelligence frameworks simultaneously.
The DOJ has the Epstein files. The ICIJ has the offshore records. The courts have the decisions. We have all of them — in one search.
Here's what cross-indexing finds that single-source searching never will.
What Cross-Indexing Means
When you search justice.gov/epstein, you search one dataset at a time. When a journalist downloads a PDF and Ctrl+F's through it, they search one document. When Congress holds a hearing, they review selected exhibits.
We search everything. Simultaneously. Across boundaries the government drew to keep you from connecting the dots.
**37 indexes. 10.9 million documents. 42 GB. One query.**
| Index | Documents | What It Contains |
|-------|-----------|-----------------|
| epstein_files | 398,525 | All 12 DOJ datasets + archive.org court records |
| icij_relationships | 3,339,267 | Every relationship edge in the Panama/Pandora Papers |
| oz_decisions | 2,060,919 | Federal court decisions |
| icij_offshore | 2,016,524 | Every offshore entity in the ICIJ leaks |
| iocs | 938,163 | Threat intelligence indicators |
| block_events | 1,060,847 | Security events |
| + 31 more indexes | ... | Blog, CISA KEVs, adversaries, phishing, pulses |
When you search one name, we hit all of them. The connections that matter are the ones that cross boundaries.
Case Study: The Trump Deep Scan
We searched "Donald Trump" and every known alias — DJT, Katie Johnson, John Barron, David Dennison, Individual-1 — across every index.
Single-source result: A pile of documents.
Cross-indexed result: **A network map that proves operational integration, not social proximity.**
Here's what only appears when you cross-index.
Cross-Index #1: The Same Email in Three Datasets
Nicholas Ribis emailed Epstein on April 25, 2016:
> "Remind ur friend that I designed the DJT business comeback and now he may be the President"
This email appears in:
- **House Oversight** (congressional exhibit format, with headers and metadata)
- **Dataset 11** (raw email extraction, with MIME encoding artifacts)
- **Dataset 10** (check images — Ribis wrote checks to Epstein, photographed with the same email threads)
One email. Three datasets. Each dataset gives you different metadata. The House Oversight version gives you the exact timestamp (2:04:49 PM). Dataset 11 gives you the email thread context. Dataset 10 gives you the **financial transaction** — Ribis was paying Epstein while discussing Trump's presidency.
A single-source search finds an email. Cross-indexing finds **a financial relationship wrapped around a political relationship wrapped around a criminal enterprise.**
Cross-Index #2: The Katie Johnson Complaint Chain
The rape complaint against Trump and Epstein (Case 1:16-cv-04642) appears across:
- **House Oversight**: Epstein forwarding "katie johnson v trump complaint.pdf" to 6 recipients
- **House Oversight**: The actual sworn declaration — "tied me to a bed... violently striking me in the face"
- **House Oversight**: The lawsuit refiled with corroborating witness (Case 1:16-cv-07673)
- **Dataset 10**: Financial compliance teams screening for "Katie Johnson v. Donald J. Trump and Jeffrey E. Epstein: Trump Child Rape Claim"
- **Dataset 10**: The same compliance review that also pulled an ICIJ Offshore Leaks Database hit on a related entity (Prytanee, LLC)
- **Dataset 9**: DOJ Office of Professional Responsibility review referencing the allegations
- **Archive.org**: Related court proceedings from the Maxwell trial
One lawsuit. Seven cross-references. The compliance team at what appears to be a financial institution was simultaneously screening the Trump rape complaint AND checking the ICIJ offshore database. They found both in the same review. **Their own internal documents prove the connection between the Epstein network and offshore financial structures.**
No single dataset gives you that. Cross-indexing does.
Cross-Index #3: The 28 Girls — Three Perspectives
George Houraney's account of the Trump-Epstein party with 28 "calendar girls" appears in:
- **Dataset 8**: Internal SDNY prosecutor email circulating the Vanity Fair article (July 10, 2019)
- **Dataset 9**: The exact same prosecutor email (duplicate across datasets — the DOJ split it)
- **Combined JP2**: OCR scan of the same email from a different imaging pipeline
- **House Oversight**: Epstein forwarding Jill Harth's sexual assault account to Michael Wolff (December 15, 2018)
The DOJ split the same email across two datasets. Without cross-indexing, you might find one copy and think you've seen everything. Cross-indexing reveals **the DOJ's own dataset boundaries are artificial** — the same evidence appears in multiple places because they broke it up.
Cross-Index #4: Mar-a-Lago Across Five Sources
Mar-a-Lago — Trump's club, documented recruitment venue, and the location where a 13-year-old was introduced to the network — appears across:
- **Archive.org**: Government Exhibit GX-823-R from the Maxwell trial — a **Mar-a-Lago Club personnel action notice** entered as evidence in a federal criminal prosecution
- **Dataset 9**: Court subpoena served on "The Mar-a-Lago Club, 1100 South Ocean Boulevard, Palm Beach, Florida 33480"
- **Dataset 8**: Duplicate of the same subpoena
- **Dataset 11**: Emails from Jonathan Farkas discussing Mar-a-Lago fundraisers with Epstein
- **House Oversight**: Trump Properties LLC correspondence on Mar-a-Lago letterhead
- **House Oversight**: Michael Wolff sharing draft book passages about Trump at Mar-a-Lago
The Maxwell trial exhibit (archive.org) connects to the Epstein emails (Dataset 11) connects to the court subpoena (Dataset 9) connects to the Trump corporate entity (House Oversight). **Mar-a-Lago isn't just a social venue in the Epstein files — it's an evidentiary throughline from Trump's corporate structure to the criminal prosecution of his co-defendant's recruiter.**
Cross-Index #5: The ICIJ Offshore Layer
When you search "Trump" in the Epstein files alone, you get 1,104 documents. When you add the ICIJ offshore database — 2 million entities from the Panama Papers, Pandora Papers, and related leaks — new patterns emerge.
The Dataset 10 compliance review that screened the Katie Johnson complaint **simultaneously pulled an ICIJ Offshore Leaks Database hit**. The document reads:
> "Etienne Binant _ ICIJ Offshore Leaks Database HIT.pdf; Katie Johnson v. Donald J. Trump and Jeffrey E. Epstein_ Trump Child Rape Cl.pdf"
A financial compliance team found offshore connections AND the rape complaint in the same screening. They filed both together. The DOJ released it. We indexed it.
The ICIJ database has 3.3 million relationship edges. The Epstein files have 398,525 documents. When you search across both, you find **the compliance teams were already cross-referencing before we were.** They knew. They documented it. They filed it. Nothing happened.
Cross-Index #6: The Social Graph
Our co-occurrence engine searches across ALL indexes simultaneously and builds a network map. For Trump, it found:
**55 connections. 596 co-documents with Epstein. 210 with Maxwell.**
But here's what only cross-indexing reveals: **the intermediaries.**
- **Michael Wolff** (75 co-docs): Information conduit between Epstein and Trump — found in Dataset 11 emails, House Oversight exhibits, and the blog index (our published analysis)
- **Nicholas Ribis** (found via DJT alias search): Epstein financial relationship in Dataset 10 (checks), political commentary in Dataset 11 (emails), and congressional exhibits in House Oversight
- **Tom Barrack** (found in Dataset 10, Dataset 11, House Oversight): Bridge between Epstein's social network and Trump's inaugural committee — later subpoenaed for financial irregularities
- **Alexander Acosta** (55 co-docs): "Belonged to intelligence" in Dataset 9 DOJ OPR report, hired by Trump in House Oversight correspondence
No single dataset reveals the intermediary network. The social graph only appears when you search everything.
Why Single-Source Search Fails
The DOJ released the Epstein files across 12 separate datasets. Each dataset was released at a different time. The House Oversight Committee released its own set. Archive.org has the court records. The ICIJ has the offshore leaks.
**They are designed to be searched separately.**
That's not a conspiracy. It's how government data works. Each agency, each committee, each release creates its own silo. The connections between silos are where the truth lives.
The DOJ compliance team that screened Katie Johnson AND the ICIJ offshore database in one review? They cross-indexed. The SDNY prosecutors who circulated the Vanity Fair article internally? They cross-referenced. The House Oversight Committee that compiled the Ribis emails? They cross-indexed.
They all did it internally. None of them made it public. None of them made it searchable.
We did.
The Numbers
| Metric | Single-Source | Cross-Indexed |
|--------|-------------|---------------|
| Trump documents | ~200 (one dataset) | **1,104** (all datasets) |
| Aliases found | 1 | **4 active** (Trump, DJT, Katie Johnson, Mar-a-Lago entity) |
| Network connections | ~5 (obvious names) | **55** (intermediary network) |
| Court cases | ~10 | **168** |
| Compliance cross-references | 0 | **ICIJ + rape complaint in same screening** |
| Frameworks applied | 0 | **5** (CARVER, DREAD, Diamond, ACH, Social Graph) |
**The difference between 200 documents and 1,104 is not volume. It's visibility.** The 904 documents you miss in a single-source search contain the intermediaries, the financial transactions, the compliance reviews, and the court records that connect social proximity to operational integration.
How to Use It
Search any name. We search all 37 indexes. The results include which dataset each hit came from, so you can trace the provenance.
The DOJ can remove photos from justice.gov. They can split datasets. They can release files on Friday afternoons. They can redact names while leaving Social Security numbers visible.
They can't un-index what we've already indexed. And they can't prevent you from searching across boundaries they designed to be unsearchable.
**398,525 Epstein documents. 7.4 million ICIJ records. 2 million federal decisions. One search bar.**
**[https://epstein.dugganusa.com](https://epstein.dugganusa.com)**
*All data sourced from government releases (DOJ EFTA), international journalism consortiums (ICIJ), public court records (archive.org), and federal decisions. No private data. No hacking. No felonies. Just government documents, made searchable, pointed at the government.*
*95% confidence cap. We present evidence, not verdicts.*
**Published by DugganUSA LLC | February 27, 2026**
**DOI: [10.5281/zenodo.17810099](https://doi.org/10.5281/zenodo.17810099)**
*Her name was Renee Nicole Good.*
*His name was Alex Jeffery Pretti.*
Comments