How can I check if a PDF is properly redacted?
Automated verification compares each page’s rendered pixels against its underlying PDF text layer. When selectable characters sit inside a bounding box whose rendered pixels show almost zero color variation, the words are visually hidden yet still extractable by anyone opening the file.
Sampling RGB pixel variance under every selectable word
For every page with selectable text, the checker renders the page onto a local canvas at up to 1.5x scale (capped at 4,096 px on the longest side) and extracts raw pixels using the MDN CanvasRenderingContext2D getImageData API. It pairs content stream glyph extraction with bounding-box mapping onto that canvas.
To avoid false negatives from thin box borders while evaluating solid patch color uniformity, the sampler trims 1 to 2 px from the outer edge of each word box. If the maximum standard deviation across the red, green, and blue channels is below 15, the word sits under a flat solid patch and is flagged as a failed redaction leak.
Inspecting unapplied Redact annotations before export
Beyond painted shapes, failed redaction detection inspects every page’s annotation dictionary for redaction marks whose subtype is Redact. Editors that use a two-step mark-and-apply workflow leave these annotations behind if a user saves too early, making it essential to identify unapplied redactions in court filings before submission.
When any unapplied mark or hidden word appears, select Redact it properly to permanently burn redaction boxes into the page. The file transfers directly in browser memory with zero network traffic.
Common ways PDF redactions leak sensitive text
Most public redaction failures happen because visual hiding looks identical to true deletion in a PDF viewer. Comparing common mistakes explains what the checker catches on each page.
| Cover-up method | Visual appearance | Underlying PDF data | Checker result |
|---|---|---|---|
| Black rectangle shape | Solid black bar | Original text operators intact | Flagged as hidden text leak |
| White box or white font | Blank white space | Selectable characters still present | Flagged as hidden text leak |
| Colored highlighter bar | Solid colored block | Words remain copyable and searchable | Flagged as hidden text leak |
| Unapplied Redact annotation | Framed or black mark | Text untouched until applied | Listed under marked but not applied |
| Flattened raster redaction | Solid black bar in image | Old page and text streams deleted | Clean pass with no hidden text |
Under the PDF Association ISO 32000-2 PDF 2.0 specification, redaction annotations only mark candidate regions for removal until a processor executes them. Relying on unapplied marks or vector rectangles leaves sensitive strings exposed to copy-paste and text extraction tools.
Even when page text passes cleanly, a document can still leak names or edits inside XMP properties or embedded attachments. Pair this check with a hidden PDF metadata and attachment scan before releasing sensitive files.
Understanding checker results on scanned and encrypted files
Not every PDF contains a selectable text stream, and not every file can be opened without credentials. Interpreting the result banner accurately prevents false confidence or unnecessary rework.
Why scanned PDFs report no hidden text layer
When a document is scanned from paper or exported as a flat image PDF, the checker displays a clean confirmation noting that the file has no selectable text at all. Because there is no character stream beneath the image, text cannot leak via copy-paste or search.
As noted in the Library of Congress PDF 1.7 Format Description, PDF files can hold pure raster page images without any character operators. However, if you previously ran OCR on a sensitive scan before painting black boxes, that OCR text layer will be caught by this checker.
Handling password-locked documents before inspection
Files protected with a user open password cannot be parsed until they are decrypted. If you need to audit an encrypted contract or court exhibit, unlock a password-protected PDF before scanning using the password you already hold.
Once unlocked, drop the decrypted copy onto Drop a redacted PDF to check. Remember to re-encrypt the verified file afterward if your recipient workflow requires password protection.
Fixing failed redactions and auditing documents offline
Testing an unredacted or partially redacted draft on a third-party server creates an immediate disclosure risk. Running verification locally keeps unreleased legal, medical, and financial records inside your own device memory.
Handing off leaked files directly to permanent redaction
When the report lists leaked pages, leaked words stay masked behind bullet characters until you check Show text. This prevents sensitive account numbers or names from appearing unexpectedly if someone is near your screen.
After fixing exposed pages in Redact PDF, you may later want to make clean scanned pages searchable after redacting if the final recipient requires searchable archives. Always run OCR only after permanent redactions are burned in, never before.
Tip: Run this checker on every PDF received from external vendors or opposing counsel as well as your own outgoing files, because unapplied redactions are easy to overlook in desktop previews.
Protecting unredacted drafts by keeping checks in your browser
Complete sanitization under the NIST SP 800-88 Rev. 1 Media Sanitization Guidelines requires verifying that purged information cannot be recovered from the output medium. Local browser checking gives you a repeatable verification step before publishing records online.
Once the page loads, every canvas render and annotation check runs in memory with zero network calls. Select Change file to clear the current report and audit another document.