Remove selected visible areas from a sharing copy
This workflow renders every PDF page into an image, replaces selected rectangles with black pixels, and builds a new PDF. It does not keep source text objects beneath those rectangles. You must identify and add each area that needs removal.
The whole document becomes image-only, including pages without selected rectangles. Searchable text, editable fields, links and accessibility structure are not preserved. Keep the original privately and confirm that this output meets the recipient's needs.
Decide what must be removed and what must remain useful
| Your situation | Recommended approach |
|---|---|
| A visible name or number must be removed from a sharing copy. | Identify every occurrence, including scans, repeated headers and context that reveals the same information. Add a rectangle for each location; the tool does not find matching text automatically. |
| You only want to hide a page edge or draw a cover box. | Cropping and ordinary overlays can leave underlying source content in place. Use this separate redaction workflow when the intention is to remove selected rendered information from the delivered copy. |
| The recipient needs searchable text, editable fields or validated signatures. | An image-only conversion may be unsuitable. Use a workflow designed to preserve the required features and have the final document reviewed before relying on it. |
Redact the sample's first-page heading
The synthetic sample's first heading provides a visible target. This exercise checks one known rectangle; it does not establish that a real document has been fully sanitised.
- Choose device-document-lab.pdf and select Load PDF pages. Under Add a black redaction area, set Page to 1.
- Set From left (mm) to 10, From top (mm) to 18, Width (mm) to 180, and Height (mm) to 12. Select Add area and confirm the rectangle appears in Redaction areas.
- Set Raster quality to 96 DPI · smaller. Read and select the image-only confirmation, then select Securely redact PDF.
- Download the redacted PDF. Inspect page 1: the heading should be covered, while the PAGE 1 OF 5 label below it remains visible. Inspect the remaining pages too.
Expected result. The output should contain five image-only pages and one black rectangle on page 1. Ordinary PDF text extraction should not recover the source text layer. This absence alone does not prove every sensitive detail was selected or removed from the visible images.
Check the downloaded result
- Inspect every rectangle edge at high zoom. Check complete characters, nearby identifiers and repeated occurrences, while confirming that information you intend to retain remains legible.
- Try searching and copying text from the output, but also inspect every page visually. Search can return no results simply because all pages are images, including unredacted information.
- Confirm page count, orientation and important form appearances. Rendering depends on what the PDF engine can display; unusual annotations or unsupported form types may not look like the source application.
- Check the file and filename you intend to send. Avoid attaching the original by mistake; other copies on your device remain untouched.
Troubleshooting
- The processing button stays disabled.
Typing rectangle values does not add an area by itself. At least one area and the image-only acknowledgement are required.
Try this: Select Add area, inspect the Redaction areas list, then read and select the confirmation. Add each further rectangle separately.
- The rectangle is misplaced or part of a word remains.
Coordinates are measured in millimetres from the visible page's top-left. Guessing from a small preview can produce an incomplete selection, and a draft is not drawn as an interactive overlay on the preview.
Try this: Return to the source, measure the target again and use a rectangle that safely covers the entire intended area. Inspect the downloaded result rather than treating coordinates as proof of coverage.
- A large file fails or the output looks softer.
Rendering every page requires memory and replaces vector content with JPEG images. Higher requested DPI uses more pixels, while device limits can reduce the actual resolution.
Try this: Try a smaller document or a lower quality setting, then recheck legibility. A worker-support error requires a browser with the rendering features this tool uses.
Limits and compatibility
- Manual selection and scope
- There is no automatic search-and-redact, OCR detection or guarantee that every identifier has been found. Pages, output bytes and area counts are limited for the device. Rectangles must fit their selected visible pages.
- New file, not forensic certification
- The worker creates a new image-based PDF with new document properties rather than transferring source text, attachments or interactive structures. Visible information outside your selections remains. This is not a forensic assurance, legal compliance assessment or deletion of originals, backups and downloads.
Your files and form values are processed on this device. Downloaded copies remain wherever you save them. See how to check local processing and our privacy policy for site data and optional services.
Common questions
- Can I undo a rectangle after creating the PDF?
- Work from the retained original to revise selections. The new page image contains black pixels in that area and does not keep the removed source objects as editable layers.
- Can I add OCR after redaction?
- You can recognise text in the resulting images, but review that new output separately. OCR does not determine whether you missed a visible name elsewhere or whether the document is safe to disclose.
Continue your document workflow
Maintained by Shakti Prasad Ram. Found a mismatch? Report the tool, browser and steps without sending a sensitive document. Back to the tool