Edit PDF tool

Redact PDF Online

Remove selected sensitive regions through a flattened, image-based export—not a removable visual overlay.

Step 1 · Upload

Upload and prepare

Add your file

Drag & drop or select a file

20 MB per PDF · up to 50 pages

Critical distinction

A black rectangle is not automatically redaction

Drawing a shape above text changes what the page looks like, but the covered text can remain inside the PDF. A recipient may remove the shape, search for the value, select it, copy it or extract the underlying object.

Always verify the downloaded document before sharing.

Visual hiding versus removal

What makes this export different

The current implementation does not copy the original PDF page objects into the result. It creates a new flattened, image-based document.

Unsafe substitute

Rectangle placed over content

  • Original text or image may remain underneath.
  • Layers, annotations or edit history may expose it.
  • Search, copy or extraction may still find it.
  • The covering shape may be removable.
Processed output

Rasterized permanent redaction

  • Each complete page is rendered at 300 DPI.
  • Selected regions are painted into page pixels.
  • A new image-only PDF is created.
  • Original page objects are not copied to the result.

Implementation

How redaction works here

  1. 01

    Render the source

    Imagick and Ghostscript render each PDF page to a 300 DPI PNG image. This flattens the visible page appearance.

  2. 02

    Paint selected regions

    Every validated page region is filled with black pixels. Regions must stay inside the page and should include a small margin around the complete sensitive item.

  3. 03

    Create a new PDF

    The processed page images are placed into a newly generated PDF with the source page dimensions. It contains no searchable text layer unless OCR is applied later.

Content types

Text, images and repeated information

Text redaction

Cover the entire word, number and surrounding glyph edges. Search the source for variations, spacing changes and repeated appearances in headers, footers or tables.

Image redaction

A selected area is painted into the rasterized page, so it can cover faces, signatures, barcodes or parts of scans. Check at high zoom for exposed edges.

Repeated values

Marking one occurrence does not find another automatically. Review every page, including duplicates, appendices and content embedded inside images.

Beyond the visible paragraph

Metadata, layers, annotations and comments

Sensitive information can exist outside the obvious page content. Secure review must consider the whole document, not only the marked rectangle.

Metadata and document properties
The export creates a new PDF and does not intentionally copy source title, author, subject or keyword fields. However, this is not a dedicated metadata inspection tool. Review the downloaded file’s properties before distribution.
Hidden layers and source objects
The final pages are rasterized rather than assembled from the original page objects, so source layers and covered text objects are not carried into the new page content.
Annotations and comments
Interactive source annotations and comment objects are not copied as interactive objects. Anything visible during rendering may become part of the page image, so inspect the output for visible review marks.
Attachments and other locations
The exporter creates a new page-only PDF rather than copying source attachments. Still check whether the sensitive value appears elsewhere in the visible document or in files you plan to share alongside it.

Common uses

Information people commonly need to remove

  • Personal identifiers and contact details
  • Account, tax and reference numbers
  • Signatures, faces and identification images
  • Confidential pricing or commercial terms
  • Health, legal or employment information
  • Names of protected or unrelated parties

Mandatory final review

Verify the exported PDF, not the preview

Always verify the downloaded document before sharing. Preserve the restricted original separately and test the exact file you intend to distribute.

  1. SearchSearch for every original value and common variation.
  2. Select and copyTry selecting the marked area and paste copied content into a plain-text editor.
  3. Inspect visuallyZoom into every redaction and check neighboring pixels, pages and repeated sections.
  4. Review propertiesInspect document properties and confirm no sensitive title, author, subject or keyword remains.
  5. Open elsewhereTest the output in another current PDF reader before release.

Processing behavior

Quality and workflow limitations

Because all pages become 300 DPI images, selectable text, hyperlinks, forms, bookmarks and other interactive page behavior are not preserved. Fine vector artwork may render differently, and the output can be larger than the source. The current limits are 20 MB per PDF, 50 pages and 1–100 redaction regions.

Troubleshooting

Fix problems before distribution

A character edge is still visible

Return to the restricted original, enlarge the region with a safety margin and export a new result.

The value appears on another page

Search the complete source and mark every repeated occurrence, including headers, footers and scanned images.

The output is not searchable

This is expected because the current export is image-only. If you run OCR afterward, repeat the full redaction verification.

The file is larger or looks different

Rasterizing every page changes its representation. Inspect small text and artwork; keep the original restricted rather than overwriting it.

The PDF cannot be processed

Check encryption, damage, the 20 MB upload limit and the 50-page maximum. Complex files can also reach server limits.

Properties still concern you

Do not distribute the file until you have inspected it with an appropriate metadata-analysis workflow.

Sensitive-file handling

Privacy and temporary deletion

Redaction requires VPS processing with Imagick and Ghostscript. Generated files become eligible for cleanup after 15 minutes and, while scheduled cleanup is healthy, are normally removed by a later run. Use an offline or organization-approved workflow when online processing does not meet the document’s sensitivity requirements.

Security Centre File Deletion PolicyPrivacy Policy

Redaction questions

Frequently asked questions

Is drawing a black rectangle secure redaction?

No. A rectangle can sit above the original text or image while the underlying object remains searchable, selectable or extractable. Use the processed Redact PDF output, not a visual overlay or preview.

How does this tool remove content?

It renders every source page to a 300 DPI image, paints the selected regions into those pixels and creates a new image-based PDF. The original page objects are not copied into the exported document.

Will text remain searchable after redaction?

No. The current export rasterizes every page and does not add a new OCR layer. If searchable unredacted text is required, apply OCR afterward and verify that redacted values do not reappear.

Does the tool remove original metadata and comments?

The export creates a new PDF rather than copying source document objects, annotations or original document properties. Still inspect the output and do not treat this tool as a dedicated metadata-audit product.

Can I redact an image or scanned PDF?

Yes. Because the final page is rasterized, a selected region is painted into the page pixels whether the visible content began as text, a photograph or a scan.

What are the processing limits?

A source can be up to 20 MB, the document can contain up to 50 pages and one job can contain between 1 and 100 redaction regions. Complex pages may also encounter server resource limits.

How should I verify the downloaded document?

Search for and try to select the original value, copy text, inspect every marked region at high zoom, review document properties and check every page for repeated information.