Redact PDF

Use the tool that matches the material on the page. Text selection snaps a rectangle to selectable words. Rectangle and ellipse cover a fixed area. Freehand follows a drawn path and offers three stroke widths. Find marks each matching word or number across the loaded document. Every mark remains editable until export, so you can move, resize, delete, undo, redo, or clear it.

• Select text when the PDF already has a text layer
• Draw a rectangle or ellipse over images, signatures, or mixed content
• Use freehand for an irregular area that a box would not fit well
• Find a repeated name, reference, or number across several pages

A mark is a proposed redaction, not proof that the whole detail is covered. Zoom in and check its edges. If a character, line, or image remains visible, enlarge the mark before applying.

Covering words with an editable annotation is not enough because the text can remain selectable underneath. Redact PDF uses a different export path. For each page with at least one mark, the browser renders the page into a bitmap, burns the black rectangles, ellipses, or freehand strokes into those pixels, and stores the rendered page as a JPEG-backed PDF page. The original text and objects from that marked page are not copied into the output.

This also means all text on a marked page becomes non-selectable, not only the text beneath the mark. Pages with no marks are copied into the new document and keep their selectable text. That tradeoff is useful to know before redacting a long report: mark only the pages that need it, and keep the source file in a secure location in case you need its text later.

Marked pages are rendered at twice the PDF point scale and encoded as JPEG images with a fixed quality setting of 0.9. There is no output-quality control in the editor. Fine print, thin lines, or compressed scans can look different after rasterization, and image-backed pages may change the file size. Compare the downloaded copy with the source at normal size and high zoom. If legibility is not good enough, keep the source private and make a new redacted copy with larger marks rather than relying on an unclear result.

Auto-detect looks for four text patterns: email addresses, phone-like number strings, card-number lengths, and US Social Security numbers. In the supplied three-page test record, each option created reviewable marks on the pages where matching patterns appeared. Find also marked repeated names across all three pages.

The detector does not know whether a match belongs to a real person or whether an unusual format is sensitive. A long reference number may look like a phone or card number, while a spaced or broken line can be missed. Treat the results as a starting list. Review every page, remove false positives, add missed details, and enlarge marks whose edges do not cover the complete text.

A scanned PDF may contain only page images, so Find and pattern detection have no text to inspect. Open Scanned, choose the languages present, and use Read scanned text. OCR renders and reads the image pages locally, then makes recognized words available to Find and Auto-detect. The supplied two-page scanned record completed OCR and email detection in the latest build.

OCR is an aid, not a verification step. Small type, blur, handwriting, unusual spacing, or a wrong language choice can produce missing or incorrect words. Compare the marks with the visible page and draw manual shapes over anything OCR did not recognize.

When the source has an open password, the drop area shows an inline unlock prompt before the editor opens. Enter the known password and choose Unlock. The password is used by the browser to open that copy for processing. Frisqoo does not upload the source file or password for processing.

The redacted download is a new PDF and does not keep the source password. If the result also needs access control, verify the redactions first and then use Protect PDF to add a new password.

Apply and download rebuilds a new PDF in page order. Marked pages are rendered and flattened with the marks burned in. Unmarked pages are copied with their text intact. The saved document clears the title, author, subject, and keywords, and identifies Frisqoo as its creator and producer. This is not a promise that every possible hidden artifact or external copy has been removed.

The filename gains -redacted.pdf, so record.pdf is offered as record-redacted.pdf. The completed panel provides Download again and, when the browser supports it, Share. Keep the original private because it still contains everything that was removed from the new copy.

Redaction is a high-consequence task, so do not rely on the black appearance alone. Open the downloaded copy in a separate reader and review it as if you were the recipient.

• Inspect every marked area at high zoom for visible character or image edges.
• Search for each removed name, email, number, and reference with the reader's Find command.
• Try selecting or copying text around each redacted area. Marked pages should behave as image pages.
• Check the page count, order, orientation, images, and unmarked pages.
• Inspect document properties if metadata matters to the disclosure.

If a detail is still visible, return to the source, enlarge or add the mark, and export another copy. Keep only the version that passes the full review.

Processing runs locally in your browser. Frisqoo does not upload the source file for processing. That statement covers Frisqoo's tool path, not the security of your device, browser extensions, download folder, backups, or any app or service you use to share the result. Store both the source and redacted copy according to the sensitivity of the document.

• Removing account and card numbers from a statement before sharing it.
• Taking names, email addresses, and phone numbers out of a case record.
• Covering signatures, faces, or ID details on scanned documents.
• Preparing a disclosure copy for a legal, public-record, or support request.
• Sharing part of a report while keeping the full source private.