Parse vectors + text layer
PDF.js inspects the operator list to identify redaction-like shapes and extract text from standard and aggressive passes.
PDF X-ray
Find hidden text under PDF redactions
Catch redactions that look safe—but still leak text. Many “redacted” PDFs still contain the original text — it’s just covered by colored boxes. PDF X-ray reveals it in seconds. Run PDF X-ray directly in your browser to highlight “redactions” that still expose selectable or searchable text—and compare the visual PDF with the actual text layer side-by-side.
Live scan status
Updates as your PDF is analyzed.
Status
Ready for a PDF.
Pages
-
Findings
-
The problem
Bad PDF redaction is easy to miss.
A black box on the page can still leave behind selectable text, searchable layers, exposed metadata, or inconsistent “clean” copies.
- Text still selectable or searchable beneath redactions.
- Hidden text layers that don’t match what’s visible.
- Metadata exposing names, IDs, or locations.
- “Clean” copies that differ from what was published.
Redacted PDF still searchable? That usually means text under black box PDF or a PDF redaction overlay.
PDF X-ray is a redaction audit tool to verify PDF redaction before you share or publish.
How it works
Three steps to a clean release.
In-browser parsing + Python rules spot leaked text before it ships.
Score overlap with Python rules
Pyodide runs overlap checks (WASM-accelerated), then filters noise like repeated chars, “redacted” markers, and date-only hits to separate real leaks from noise.
Overlay + compare the clean layer
Findings draw SVG overlays, and a text-only PDF is rendered side-by-side for visual vs. extractable comparison.
Live demo
Run an X-ray in your browser.
Upload a PDF and review leaks instantly. Nothing leaves your machine by default.
Findings
- No PDF loaded yet.
Rectangles detected
-
Text items detected
-
If rectangles = 0, the PDF likely uses rasterized or non-vector redactions.
Preview
Upload a PDF to see detections and overlays.
Key benefits
Value props teams care about.
Catch failed redactions in seconds
Identify visual-only redactions where text still exists.
Side-by-side proof you can share
Export a report of findings for Legal, Security, or Comms.
Runs where your data is
Browser-first checks reduce data movement and friction.
Built for real-world PDFs
Messy scans, mixed layers, odd fonts, and legacy exports.
Product demo
See leaks instantly.
The X-ray View
- Page preview with leak highlights
- Text-layer extraction panel
- Findings list with severity
Report View
- Summary (pass/fail)
- Findings per page
- Suggested remediation steps
Add screenshots or GIFs
Place 2–3 visuals here for the demo.
What we detect
Common redaction failure signals.
- Text still present under drawn rectangles or overlays.
- “Whiteout” redactions that leave searchable text.
- Inconsistent text layer vs visual render.
- Copy/paste leaks (selectable regions).
- Suspicious metadata fields (author, title, toolchain).
- Duplicate or alternate text content patterns.
Note: PDF X-ray helps you find likely leaks fast. Final validation depends on your redaction process and toolchain.
Use cases
For teams shipping sensitive PDFs under real deadlines.
From legal to incident response, PDF X-ray keeps redaction quality high when timelines are short.
- Legal & Compliance verify filings and disclosures.
- Journalism / Research protect sources and IDs.
- Security / IR catch leaks before sharing.
- Procurement / Sales Ops scrub customer data.
Why browser-based?
Because redaction checks happen under time pressure.
- No install. Open, drop a file, scan.
- Works on locked-down devices.
- Great for quick triage and one-off checks.
Security & privacy
Your documents are sensitive. We treat them that way.
- In-browser scanning option (where possible).
- Optional “no retention” mode.
- Encrypted transport.
- Clear data deletion controls.
FAQ
Common questions before release.
Does PDF X-ray “break” encryption or passwords?
No. You can scan only what you can open or decrypt legitimately.
Will it work on scanned PDFs (images)?
If there’s no text layer, leaks are less likely—but we can still flag metadata and structural oddities.
What should I do if it finds a leak?
Re-redact using a true redaction tool that removes hidden text from redacted PDF exports, then re-export and re-scan.
Can I run it offline?
If you ship a fully local build (PWA or desktop wrapper), yes.
Is this a redaction tool?
It’s a verification tool: “Did this redact correctly?” not “Let me redact.”