FileGizmo PDF tools

Redact a PDF so the words are really gone

Drag a box over anything private and the text inside it is deleted from the file, not covered with a rectangle. Free, and your PDF is never uploaded.

Never uploadedYour file stays on this device

Unlimited edits. No daily tasks. Your work survives a crash. Your file still never leaves.

Drop your PDF here or click to choose a file PDF file · processed on this device

Why this tool

Redaction that removes the words instead of hiding them

Redaction has one job: the information has to stop existing in the file. A rectangle drawn on top of a name does not do that. The rectangle is simply the last instruction on the page, and the name is still sitting in the content stream a few lines above it, where copy and paste, a search, or any text extractor will find it in seconds. Law firms and government departments have published documents redacted that way and had the hidden text read back to them the same afternoon.

This tool works the other way round. Drag a box over the words you want gone, and on apply every character whose position falls inside that box is removed from the page’s content stream. What goes in its place is a numeric spacing instruction, so the words that survive on the same line keep the exact position the original typesetter gave them instead of sliding left into the gap. Only after the characters are gone is the black rectangle drawn, and it is there to show a reader that something was removed, not to do the removing.

Because a promise like that is worth nothing unless it is checked, the tool checks it. Once the redacted document has been written, it is opened again and its text is read back with the same library an opponent would point at it. If any character still falls inside a region you marked, that page is not released. It is rendered to pixels instead, which destroys everything underneath the box, and the result names the page so you know what happened and what it cost.

Scanned pages are the honest limit. A scan is a photograph of a document; the words on it are pixels, and there is no text layer to delete from. The tool recognises that a picture rather than searchable text sits under your box, flattens that page to an image so the pixels themselves are destroyed, and tells you plainly. It does not draw a rectangle over a photograph and report a clean redaction, which is the failure mode that makes redaction tools dangerous.

Metadata is cleared in the same pass. Title, author, subject, keywords, producer, creator, and the XMP packet are all emptied, because an author name or a revision title outside the page is a leak that survives even a perfect redaction of the page itself.

Two things this cannot do, said plainly. Vector line art under a box is covered rather than deleted, so if what you are hiding is a drawing or a hand-drawn signature rather than text, tick the flatten option before you apply. And a document that is already encrypted has to be unlocked first, because a file this tool cannot read is a file it cannot clean.

Before you send the result, do the check yourself. Open the downloaded copy in a different reader, search for every name and number you removed, select all the text on each redacted page, and paste it somewhere plain. A redaction you have verified is worth more than one you were told about.

Simple by design

How to redact a PDF on this device

  1. 1

    Choose an unlocked PDF

  2. 2

    Drag a box over every name number or line that must go

  3. 3

    Apply and download the redacted copy

Built for the whole job

Everything you need

The characters leave the file

Each glyph inside a box is cut out of the page's content stream and replaced with a spacing instruction, so the surviving words on that line stay where they were. Copy, search, and text extraction find nothing where the box is, because there is nothing left to find.

The result is checked, not assumed

After the removal the finished file is reopened and read back with the same extractor an opponent would use. If a single character is still sitting under a box, that page is destroyed as pixels instead, and the result tells you which pages and why.

Metadata goes with it

Title, author, subject, keywords, producer, creator, and the XMP metadata packet are cleared in the same pass. Author names and revision titles survive naive redaction and have exposed more than one filing.

Nothing is uploaded, and you can verify it

Processing runs on this device. Open the browser’s Network panel while the tool works and you will see no file upload request to FileGizmo. Temporary previews use local browser URLs, and closing or resetting the page releases them instead of leaving a server copy behind.

No limits, no watermarks

Use the tool without a daily task counter, account wall, output watermark, or artificial upload cap. FileGizmo does not meter a transfer it never receives. Available memory, processor speed, browser canvas limits, and the file format itself set the honest practical boundary.

Works on any device, and offline after first load

Use a current browser on desktop, Android, iPhone, or iPad without installing an app. Once this tool and its required assets load, the cached workflow works without a connection. Large jobs are usually more comfortable on a desktop with additional memory.

Good to know

Frequently asked questions

Is this different from drawing a black box?

Completely. A black rectangle is one more instruction added to the page, and the text underneath is still in the file where any extractor will read it. This tool deletes the characters from the content stream first and then draws the rectangle, so the box marks what was removed rather than hiding what was not.

How do I know the text was actually removed?

The tool checks its own work. Once the redaction is written, the finished document is reopened and its text is read back; a page with anything left under a box is flattened to an image rather than released. You should still verify yourself, by opening the download and searching for the phrase you removed.

What happens on a scanned page?

A scan holds a picture of text, not text, so there is nothing in the text layer to delete. The tool detects this rather than pretending otherwise, flattens that page to an image so the pixels under the box are destroyed, and names the page in the result. That page stops being searchable.

Does the rest of the document stay searchable?

Yes, on every page that did not need flattening. Untouched pages are not re-rendered at all, and a redacted page keeps every character outside the boxes as real, selectable text. Only pages that could not be cleaned any other way become images.

Is my PDF uploaded?

No. The file is read, previewed, redacted, and saved in your browser. Open the Network panel while the tool runs and you will see no upload. That matters more here than anywhere else on the site, because the documents people redact are the ones they can least afford to send to a stranger's server.

Learn more

Related guides

PDF guideHow to properly redact a PDFLearn why drawing a black box does not remove PDF data, what true redaction must destroy, and how to verify a redacted document before sharing it.