Anonymize sensitive areas of a photo — faces, plates, documents — with a brush or automatic face detection. Everything runs on your device.
Your photos are never uploaded and never leave the device. The face- detection model and runtime are bundled in the app; no network is used while working. There is no tracking, no analytics, no account, no ads.
iOS asks for permission only when you tap Load (Photo Library / Camera) or Save (Save to Photos). The app accesses nothing else. Full policy: Privacy Policy.
Note: pixelation is not cryptographically secure — for stronger anonymization use a coarse pixel size and the Secure scramble slider.
DeFace is open source under the MIT License. © 2026 Ben Pohl.
Source: github.com/BenPohlBasel/deface
· onnxruntime-web — face-detection runtime (WASM). MIT License, © Microsoft.
· YuNet face-detection model (OpenCV Zoo) — MIT License.
· Capacitor by Ionic — MIT License (iOS app shell).
The original image is kept at full resolution for export; editing happens at a working resolution to stay fast. The optional face detector (YuNet via onnxruntime-web) only suggests regions — refine them with the brush or erase.
Built with care · BIAS.CITY