Auto Blur Screenshot

Automatically detect and blur faces, phone numbers, emails, and QR codes in screenshots — all in your browser, nothing uploaded anywhere.

🖼️
Drop a screenshot here or click to browse
PNG, JPG, WEBP — processed locally, never sent to any server
Initializing…
Load an image and click Auto-detect, or draw boxes manually.

✏️ Draw a box on the image to manually mark a region. Click a region in the list to highlight it.

Detected regions

No regions yet. Click Auto-detect or draw manually.

How it works

Every step runs inside your browser tab using WebAssembly and JavaScript models — your screenshot never leaves your device.

😶 Face detection Uses a lightweight ONNX face-detection model via Transformers.js (WebAssembly). Finds bounding boxes for every visible face in the image.
📱 Phone & email OCR Tesseract.js runs full OCR on the image, then regex patterns scan the text for phone numbers (E.164 / local formats) and email addresses. Matching word bounding boxes are flagged.
⬛ QR code detection jsQR scans the pixel data for QR code finder patterns. When a QR code is found its bounding quad is highlighted for blurring.
✏️ Manual boxes Click and drag anywhere on the image to mark additional sensitive regions — credit card numbers, names, addresses, or anything else you spot.

Once you confirm which regions to blur, choose Pixelate (mosaic effect) or Solid fill (black bar), then click "Apply blur & download" to save the redacted PNG.

Frequently asked questions

Is my screenshot uploaded to a server?
No. Everything runs in your browser using WebAssembly and JavaScript. The image pixel data never leaves your device. The only external network calls are one-time downloads of the AI model weights from a CDN — those are anonymous and contain no image data. You can even disconnect from the internet after the models load and the tool will still work.
Why does the first scan take a while?
On the first use, the browser downloads the face-detection model (around 1–2 MB) and the Tesseract OCR data file for English (~10 MB). These are cached by your browser, so every subsequent scan on the same device is much faster — usually a few seconds.
What if a face or phone number is missed?
Auto-detection is not perfect. Small, partially occluded, or stylized faces may be missed, and unusual phone formats may not match the regex. That is why the tool lets you draw manual boxes: click and drag on the image to cover anything the auto-detector missed. Manual regions behave identically to auto-detected ones and are included in the download.