Choose the document carefully
This redactor helps prepare a copy of a PDF for sharing when specific visible areas contain information you want to remove. Select a local PDF no larger than ten megabytes, with at most ten pages. The original stays on your device. File content is processed in your browser; it is not submitted to this website as an upload. A PDF engine and a PDF builder load from this website as application assets. Use a supported modern browser and keep the tab open until the export finishes. Password-protected, damaged or unusual documents may fail to render. Keep your original separately and never treat a partially loaded preview as a completed redaction.
Mark every sensitive area
After loading, use Previous page and Next page to inspect the entire document. Drag a rectangle with a mouse or touch pointer over each sensitive area. A small safety margin is painted around each rectangle. For keyboard access, enter Left, Top, Width and Height as percentages of the page, then choose Add rectangle. The rectangle must fit within the page and have positive width and height. Undo removes the last rectangle on the current page. Clear this page removes all its rectangles. Neither control changes another page. The counter shows the total marked rectangles, which helps you notice an accidental reset before export.
Understand what the output contains
The export does not place removable black annotations over the original PDF. It renders every page into a new image, paints the marked pixels black, and builds a fresh PDF from those images. The original PDF objects are not copied into the output. This also discards searchable text, hidden text layers, links, interactive fields, original attachments and original metadata. A scanned source is treated the same way. The new file contains ordinary raster page images, so it can be opened by normal PDF readers. The visible content outside your boxes remains in those images. If confidential information appears elsewhere, it still needs another rectangle.
Check quality and document features
Pages are rebuilt at one and a half times their PDF point size, approximately 108 dots per inch. This balances browser memory and legibility for typical text documents. Extremely large pages are rejected rather than allocating unlimited memory. Small print can lose clarity, and a raster PDF may be larger than its original. Digital signatures will not remain valid in a reconstructed document. Forms become flat pictures and accessibility text structure is lost. Do not use this workflow when a recipient requires a signed original, selectable text, preserved form fields or an accessible tagged PDF. Review requirements before deciding that an image-based copy is suitable.
Read verification without overinterpreting it
Before offering a download, the tool opens the rebuilt bytes with the PDF engine and checks every page for extractable text. A successful check confirms that no searchable text was found in this generated file. It does not prove that every confidential item was covered visually. A visible name, photo, barcode or signature outside a rectangle can still disclose information even though text extraction returns nothing. The report records page count, rectangles per page, file size and resolution. Compare these with your intended document. Download the PDF and zoom into every marked region and nearby lines in an independent reader before distributing it.
Avoid common redaction mistakes
Look beyond obvious account numbers. Repeated headers, footers, QR codes, screenshots, marginal notes and appendix pages may repeat the same information. Cover the entire relevant area, including labels if their presence is sensitive. Removing text does not anonymise a face or make contextual clues disappear. Changing the selected file clears the previous working state and download link, so each document starts independently. The website does not keep a history of your redaction rectangles. Save only the reviewed output you intend to share and retain the source where appropriate. For highly sensitive material, follow your organisation’s approved review process and document handling policy.
FAQ
Is my PDF uploaded?
No. PDF content is processed locally in the browser. Application assets load from this site.
Can hidden original text be selected?
The rebuilt output contains page images rather than original text objects. An automatic extraction check runs before download.
Are signatures preserved?
No. Rebuilding discards digital signatures, forms and links.
Does verification catch missed names?
No. You must visually inspect every page and mark every sensitive visible area.
Can I use keyboard controls?
Yes. Percentage fields and Add rectangle provide an alternative to dragging.
What limits apply?
Ten megabytes, ten pages and a rendered-page memory limit. Password-protected files may fail.
Review before use
Start with a small example and check that every field represents the same document, month or service period. Save the original source separately. Review each generated result before sending it to another person. A downloaded file records the assumptions you entered and the calculation method shown on this page. It does not replace an independent document review, a payroll statement, an employment agreement or an official determination. If an input is uncertain, confirm that input before treating the displayed result as a reliable planning figure.
Related tools
Sources: PDF.js documentation · PDF-lib documentation