More free image tools
Same quality bar. Still free. Still no upload.
Your files never leave your device
Free to use, with no account and no watermark. The first run downloads the model files once and your browser caches them; the model then runs on your device, and the file you process is never sent anywhere.
This is not a promise about how carefully we handle your upload. There is no upload. The file you choose is read into memory by your own browser, processed there, and handed back as a download — the tool has no server component and makes no network request with your data. The only download is the model itself, which contains nothing of yours.
- Nothing is stored. We never receive the file, so there is no copy to retain, no retention period to disclose, and nothing to delete on request.
- Nothing is transmitted. No upload endpoint, no analytics payload carrying file contents, no third-party processor.
- You can verify it. Run the tool once so the model is cached, then disconnect from the internet and run it again. It still works, because everything it needs is already on your machine.
- Nothing persists after you close the tab. Files are held in page memory only — not in local storage, not in a cache, not in a database.
A note on wording, because it matters: we do not describe these tools as “encrypted.” Encryption protects data that travels to somebody else's computer. Your file does not travel, so there is nothing to encrypt and nothing to intercept — which is a stronger guarantee than encryption, not a weaker one. The page itself is served over HTTPS like the rest of the site.
This makes the tools safe for material you are contractually barred from uploading to third-party services — client footage under NDA, unreleased campaign assets, anything covered by a confidentiality clause.
How it works
1. Add screenshots or photos
Drop JPG, PNG or WebP files, pick them, or paste a screenshot with Cmd/Ctrl+V. Up to 10 images, 40 MB each. The files stay in this tab.
2. Find sensitive info
Florence-2 base (fine-tuned), Microsoft's model in the onnx-community ONNX build under the MIT licence, reads each line of text and where it sits. It runs on WebGPU when the browser has it and WebAssembly otherwise. The first run downloads it once (about 360 MB on WebGPU, about 230 MB on WebAssembly), and Image to Text reuses it. Tall screenshots are read in overlapping parts. Pattern checks then look for emails, phone numbers, card numbers (Luhn-checked), IBANs (checksum-checked), US SSNs, dates of birth, links, IP addresses and street addresses or postcodes. YuNet (OpenCV Zoo, MIT licence, 0.2 MB) finds faces. Find names too is optional: it adds BERT-base NER (dslim/bert-base-NER in the Xenova ONNX build, MIT licence, about 110 MB, run on WebAssembly). It is trained on English news, so other languages and all-caps text are weaker.
3. Review every box
Each find is a box coloured by category, and the list shows the text with most of it hidden. Switch single finds or whole categories off, delete boxes, or drag to draw your own. Nothing is hidden until you export.
4. Export
Solid black is the default. Blur and pixelate are also offered, set coarse on purpose. The boxes are drawn into the pixels and the image is re-encoded as a new PNG, which drops EXIF and location data. Download one PNG or a ZIP of the batch.
Questions
Will it find everything I need to hide?
No, so check every image yourself before you share it. The text reader can drop or misread small, light, stylised or angled text. A misread digit makes a card number fail its checksum, and then it is not flagged. Names are only found with Find names too on, and that model is trained on English. Addresses are only flagged when they have a house number and a street word, or a UK, US, Indian or German postcode shape. Anything missed can be boxed by hand.
Is my screenshot uploaded?
No. Reading, detection and export all run in this browser tab. The only network requests are the one-time model downloads: Florence-2 and the optional name model from Hugging Face, and the 0.2 MB face finder from Versely's file host. Your image is not part of those requests.
Can the black boxes be removed afterwards?
No. The boxes are not a layer on top of the image. The tool paints over those pixels, encodes a new PNG from the result, and saves only that file. The original pixels under a box are not in the download.
Is blur or pixelate safe?
Less safe than solid black. Light blur or fine pixelation over text can sometimes be reversed when the font is known. Here both are coarse on purpose: text boxes keep only one or two blocks per line, so they read as smudges. For card numbers, passwords and ID numbers, use solid black.
Which models does it use, and what are the licences?
Florence-2 base (fine-tuned) by Microsoft, onnx-community ONNX build, MIT licence, reads the text. YuNet 2023mar from the OpenCV Zoo, MIT licence, finds faces. The optional name finder is dslim/bert-base-NER in the Xenova ONNX build, MIT licence. All three allow commercial use.
Why is the first scan slow?
Florence-2 downloads once: about 360 MB on browsers with WebGPU, about 230 MB on the WebAssembly fallback. After that it loads from the browser cache. A screenshot takes a few seconds on WebGPU. A tall phone screenshot is read in two to four parts, and without WebGPU each part can take much longer. The name finder adds about 110 MB the first time you turn it on.
Does it keep the photo's location data?
No. The export is a new PNG drawn from the pixels, so EXIF, GPS and other metadata from the original file are not copied. The download is always a PNG, even when you add a JPG.
Related tools
Stay in the same job cluster. Free tools stay on-device; paid tools are labelled as such.
These rearrange files. Versely makes new ones.
Everything on this page works on a file you already have, which is why it costs nothing. Generating something that did not exist (video from a prompt, a voice, a score) runs a model, and that costs credits.