Local by default
If a feature can run in the tab, it does. There is no upload endpoint, no processing queue, and no temporary bucket to trust.
About
Redact is a free, browser-based redaction studio built on one rule: the file you are trying to protect should not be uploaded in order to protect it.
0
Images uploaded
0
Accounts required
0
Trackers or analytics
100%
Processing in your browser
Why it exists
Screenshots are how work gets explained now — a support thread, a receipt, an ID, a dashboard, a chat with a delivery address in it. The detail that makes the screenshot useful is often the same detail that should not travel with it.
Most redaction tools start by asking you to upload the exact file you are trying to protect. That never sat right. A tool that removes sensitive details should not collect them first, and a promise to delete the copy tomorrow is still a copy today.
Redact flips the order. Your browser decodes the image, draws it to a local canvas, and bakes the redactions into a new file. There is no upload endpoint. Smart detection downloads an OCR engine and an English model, but your image is not part of that request.
We are not anti-cloud. We are against the unnecessary step in the middle. When the browser can do the job, the safest place for the file is the place it already is.
Principles
These are not aspirations on a wall. They decide what gets shipped, what gets disclosed, and what gets left out.
If a feature can run in the tab, it does. There is no upload endpoint, no processing queue, and no temporary bucket to trust.
Faces, QR codes, and unusual names still need a manual box. A clean verification scan is evidence, not a guarantee.
Privacy and legal pages should be readable without a translator. Where a term is unavoidable, it gets an explanation.
No sign-in, no email wall, no usage history. The editor does not know who you are, and there is nothing to export about you.
No watermark, no export credit, no paid upgrade for the safe option. Ink black — the strongest style — costs nothing.
The Network tab is the receipt. The privacy page explains exactly what to check and what you should see when you check it.
How it works
The whole workflow fits in one tab. The only optional download is the OCR engine used by Smart detect.
01
Drop or paste a file
The browser reads the image locally. PNG, JPG, WebP, GIF, BMP, and AVIF are supported.
02
Cover what matters
Smart detect suggests likely PII. Draw over faces, codes, and anything the scan misses. Ink is the default.
03
Verify and export
Re-scan the flattened image, compare it with the original, then download a metadata-free PNG, JPG, or WebP.
Under the hood
The site is static HTML, CSS, and JavaScript. There is no backend that receives images, no analytics SDK, and no advertising code. The pieces that do the work are ordinary web APIs and open-source software.
Smart detection is the one feature that needs a network request after the page loads. It fetches Tesseract.js and language data; the image stays in the tab.
Honest limits
A redacted image does not prove that a document meets a privacy law, a court order, or an internal policy.
The guides explain practical redaction. They are not a substitute for a lawyer or a data protection officer.
There are no layers, text annotations, or batch exports yet. It is a focused tool for covering things.
OCR is best-effort, and a box only covers what it covers. Always look at the exported file before sharing it.
Those limits are the reason verification and a final visual check are part of the workflow, not an optional extra.
Use it for what it is
The editor is free, local, and open to inspection. Load an image, cover what should not travel with it, and check the export before you send it.
Free · No account · Nothing uploaded