Hidden Text in Court Documents: A Simple Guide

How to recover redacted information from Jeffrey Epstein court filings

Updated: April 23, 2026
For: Journalists, researchers, and the general public

What We Found

Thousands of court documents in Epstein-related cases contain hidden text that anyone can recover with simple copy and paste.

The U.S. Department of Justice published court filings online with faulty "redactions" — black bars that were supposed to hide sensitive information. The text is still there underneath, and you can see it by copying the blacked-out sections and pasting them elsewhere.

Real Example

Here's what we recovered from a major lawsuit document that was supposed to be redacted:

"signed Foundation account checks for over $400,000 made payable to young female models and actresses, including a former Russian model who received over $380,000 through monthly payments of $8,333"

That text was completely blacked out in the PDF viewer, but copying and pasting revealed the hidden details.

How This Happened

The Problem

When creating a PDF redaction, there are two ways to do it:

❌ Wrong Way (What DOJ Did Initially):
- Type the document normally
- Draw black rectangles on top of sensitive text
- Result: Text is hidden visually but still exists in the file

✅ Right Way (What DOJ Does Now):
- Print the document to an image (like taking a photo)
- Run text recognition software on the image
- Result: Hidden text is completely destroyed

The Timeline

• December 2025: DOJ published thousands of court documents with faulty redactions
• February 2026: Someone (possibly journalists) discovered the problem
• February 25, 2026: DOJ quietly replaced all files with properly redacted versions
• Today: Original files are only available through Internet Archive backups

What Documents Are Affected

We've identified over 12,000 court filing PDFs with this problem across major Epstein cases:

Most Revealing Documents

Across our analyzed catalog of 7 cases (56 documents, 280 pages with redaction bars):

• 719 recoverable text fragments confirmed after cross-validation
• Content includes: Financial transactions, entity names, operational details, payment amounts, privilege-log email addresses, deposition names

How we checked our work

Every finding was cross-validated using two independent methods:

1. Pixel-darkness detection — our main tool, which renders each page as a high-resolution image and looks for words whose position is covered by a dark rectangle.
2. PDF structural detection — an independent open-source tool (Lee Drake's unredact) that inspects the raw PDF drawing commands to find filled black rectangles.

The two methods agree on the core findings but catch different failure modes: the pixel method catches redactions burned into scanned-document images (which the structural tool cannot see), and the structural tool catches narrow inline bars thinner than a whole word (which the pixel method occasionally misses). When the two disagreed, we visually inspected the pages — which caught one incorrect rejection (restored) and confirmed the 21 entries we dropped as non-evidentiary (PACER headers on fully-sealed pages, FOIA exemption labels, literal "xxxx" placeholders).

How to Access These Documents

Step 1: Find the Original Files

The current DOJ website has fixed versions. You need the Internet Archive (Wayback Machine) versions from before the fix:

Go to: web.archive.org

Search for: justice.gov/multimedia/Court Records/[CASE NAME]/

Use dates: December 2025 through February 2026

Step 2: Download Specific Documents

Example URL pattern:

https://web.archive.org/web/20251228132625/https://www.justice.gov/multimedia/Court%20Records/Government%20of%20the%20United%20States%20Virgin%20Islands%20v.%20JPMorgan%20Chase%20Bank,%20N.A.,%20No.%20122-cv-10904%20(S.D.N.Y.%202022)/001-01.pdf

Replace the case name and document number with what you're looking for.

Step 3: Test for Hidden Text

1. Open the PDF in any viewer
2. Find sections with black bars (redactions)
3. Click and drag to select the black area
4. Copy (Ctrl+C or Cmd+C)
5. Paste into a text editor (Ctrl+V or Cmd+V)

If you see text appear, the redaction was faulty!

What Kind of Information is Hidden

Financial Details

• Specific payment amounts to individuals
• Bank transfer details
• Entity ownership structures
• Loan arrangements

Operational Information

• Names of previously unknown entities
• Business relationships
• Property transactions
• Legal strategies

Investigation Details

• Evidence handling procedures
• Witness information (redacted for privacy)
• Law enforcement communications

Important: We focus on systemic patterns and public interest information, not individual privacy details.

Tools for Researchers

For those wanting to do systematic analysis:

Simple Method (No Technical Skills)

1. Manual copy/paste testing on individual documents
2. Search for key terms like entity names, dollar amounts
3. Document findings in spreadsheets or notes

Advanced Method (Some Technical Skills)

We've created tools that automatically:
- Scan hundreds of PDFs to identify which have faulty redactions
- Extract all hidden text from documents
- Compare original vs. fixed versions

Tools available at: Contact researchers or check public repositories

Legal and Ethical Notes

This is Legal Research

• These are public court documents
• The vulnerability was created by the government agency
• We're using standard document analysis techniques
• Responsible disclosure focuses on systematic transparency, not individual privacy

Responsible Use

• Protect victim privacy — don't republish personal details
• Focus on institutional accountability — banks, law enforcement, estate management
• Verify information against other sources when possible
• Consider public interest vs. individual harm

Why This Matters

Government Transparency

This reveals significant problems with:
- Document security practices in federal agencies
- Transparency vs. secrecy in high-profile legal cases
- Technical competency in handling sensitive information

Historical Record

The hidden information provides unprecedented detail about:
- Financial networks supporting Epstein's operations
- Institutional relationships with banks and service providers
- Investigation and prosecution strategies

Accountability

Recovered information may reveal:
- Previously unknown entities and relationships
- Financial flows that weren't disclosed in testimony
- Coordination between defendants and service providers

Getting Started

For Journalists

1. Start with high-priority cases (Giuffre v. Maxwell, USVI v. JPMorgan)
2. Focus on financial documents (highest recovery rates)
3. Cross-reference with existing reporting for context
4. Verify findings through multiple sources

For Researchers

1. Download our case inventory to prioritize targets
2. Use systematic scanning rather than random sampling
3. Document methodology for peer review
4. Share findings through established research channels

For General Public

1. Start with our verified examples to understand the technique
2. Pick specific topics you're interested in (finances, real estate, etc.)
3. Work in groups to cover more documents efficiently
4. Report significant findings to journalists or researchers

Frequently Asked Questions

Q: Is this hacking or illegal?
A: No. These are public court documents. We're using standard copy/paste functionality that works in any PDF viewer.

Q: Why didn't DOJ fix this earlier?
A: PDF redaction is technically complex. Many organizations make this mistake. DOJ fixed it once discovered, but the originals are preserved by Internet Archive.

Q: How much hidden text is there?
A: From our sample, about 50-60% of court filings have some recoverable text. Some documents have hundreds of hidden words.

Q: How do you know the recoveries are real and not software bugs?
A: We verified every entry with two independent methods (pixel-darkness rendering and PDF-structure inspection using an open-source tool built by a separate researcher), and visually spot-checked pages where the two methods disagreed. The 21 entries our first pass flagged but couldn't substantiate were dropped from the published catalog.

Q: Can I get in trouble for doing this?
A: These are public records with a technical flaw. Standard document research is protected activity. Use common sense about republishing sensitive personal information.

Q: Are other cases affected?
A: We've focused on Epstein-related cases, but the same technical issue likely affected other DOJ multimedia archives during the same time period.

Contact

For questions about methodology, tool access, or collaboration:
- Technical questions: See detailed technical report
- Media inquiries: Standard research disclosure practices
- Academic collaboration: Documented methodology available for peer review

This guide demonstrates standard document analysis techniques for transparency research. All methods described use publicly available court records and standard software functionality.