PDF Redaction in the browser is not a generate
/free-tools/pdf-redact never spends a credit. Do not pay a model to do an in-browser file job.
PDF Redaction never spends a credit. Do not pay a model to do an in-browser file job.
Versely has no free tier for generation. Every model call costs credits. This tool is free because it never calls a model, an API, or the network. It draws boxes on a rendered PDF page, fills them in a bitmap, and rebuilds the file from those pixels. The glyphs do not survive. That is redaction. Asking Veo — or any catalog row — to "black out the contract" is a generate you will still be billed for, and it will not be redaction.
Raster, not generate
A black rectangle sitting on live text is the classic failure. The letters stay in the content stream. Anyone can delete the rectangle or copy the hidden run. This tool rasterises the page, burns your boxes into the bitmap, then writes only that image. Underlying text is gone, not merely covered.
The result is a PNG-backed page, not a JPEG. JPEG ringing at the edge of a black box can carry faint traces of the original pixels. PNG stores the filled region as flat black. The file is larger. That is the cost of not leaking what you just covered.
Nothing is uploaded. The file is read into memory on your machine, processed there, and offered back as a download. It keeps working with the network off. No account. No watermark. No credit meter.
That is the whole product. It is not "AI redaction." It is boxes into pixels.
Credits are for models
Generation on Versely always costs credits. The cheapest catalog job is still not zero. There is no daily free allowance and no 0-credit generate. Subscriptions and packs top up a balance; unused credits roll over while the subscription is active. Every generate shows its credit cost before you confirm.
None of that arithmetic applies here. PDF Redaction is in the free-tools set: Canvas and the File API, on your device. Treating it like a model — picking a row, writing a prompt, confirming a charge — is how a one-page NDA becomes an image-to-video bill and a PDF that still copies.
Do not:
- Upload the PDF to an image model and prompt "censor the names."
- Generate a video of a black bar moving across a page.
- Ask the agent to "redact this" with a video or image endpoint.
Those are generates. They spend credits. They leave text in the file or invent a picture of a file. Draw the boxes. Rebuild.
The cost page is a different surface
/cost is the credit-formula hub: per second, per 1,000 characters, per megapixel, per tier, per job. It exists so you can see which meter a real generate will hit before you dispatch it. Redaction does not appear there because it has no meter.
If the job is "remove this paragraph from a PDF so it cannot be copied out," you are on /free-tools/pdf-redact. If the job is "make a 5-second clip of a document," you are on a video row and you will pay that row's rate. Name the object. File transform versus generate. Only one of those is free, and only because it never touches a model.
FAQ
Is PDF Redaction actually free?
Yes. No account, no limit, no credits. It costs nothing to run because the work happens on your device. Generating with an AI model on Versely does cost credits; there is no free tier for generation.
Do my PDFs get uploaded?
No. Browser, Canvas, File API. The file stays on your machine. That is also why it works offline.
Why not draw a black box in a normal PDF editor?
A box is another drawing command on top of the text. The glyphs remain. This tool burns the boxes into a raster and writes the image. Copy-out fails because the letters are not in the file.
Should I send the PDF to a video model if I also need a social clip of the redaction?
Redact first, in the browser, for free. The redacted file is the source of truth. A social clip is a later generate, billed in credits, and it is a picture of a page — not a substitute for redaction.