About this tool
Blur bystander faces detected in your photos, in batches.
Batch Face Privacy Blur uses the browser's built-in FaceDetector API to locate faces in a photo you own, paints an 18-pixel Gaussian blur over each detected bounding box on a canvas, and re-encodes the result as a PNG — which also drops the original EXIF block, including any GPS coordinates. It reports how many faces it blurred, so you can see whether detection found what you expected before you share the file. It is for anyone posting a photo where bystanders, children, or colleagues should not be identifiable.
Open Batch Face Privacy Blur on AltFTool — it loads instantly in your browser.
Choose the photo you own with the 'Local file(s)' picker; if you select several, the panel names the one being processed — only the first file is used.
Press 'Run local workbench' and the browser's FaceDetector locates faces, with an 18px blur filter drawn over each detected bounding box on a canvas in your tab.
The canvas is re-encoded and downloaded as altftool-batch-face-privacy-blur.png, and the 'Verified result' panel reports width, height, 'PNG without original EXIF block' and how many faces were blurred — or that no face was found.
Detected regions are re-drawn through a canvas blur filter, so the photo still reads naturally instead of being punched full of rectangles.
Because the output is re-encoded to PNG from canvas pixels, the original EXIF block — including GPS coordinates — is not carried over.
The result reports the face count, and says so explicitly when no face was found or when your browser has no FaceDetector, rather than handing back an untouched image that looks processed.
Only those that implement the experimental FaceDetector API — in practice Chromium-based browsers, and often only with the relevant flag enabled. If it is unavailable the image is still re-encoded and stripped of EXIF, but the report will say no faces were blurred, so check that line every time.
No. The blur is applied to the pixels and then the image is re-encoded, so the original detail is not stored anywhere in the output file. Keep your own copy of the unblurred original if you need it.
No. Detection, blurring and encoding all run in your browser tab, and the blurred PNG is produced as a local download. The image never goes to a server.
No — always check the output yourself. Automatic detection misses faces in profile, at small scale, in poor light, or partly hidden, and it does not touch other identifying details such as name badges, licence plates, tattoos, or reflections.