AI Image Editor — Background Removal, Blur, Face Anonymize & Super Resolution
AI Image Editor
Edit images with AI locally in your browser — remove background, blur background, replace background, face blur anonymization, super resolution. All processing happens on your device, no file upload.
📷Drop image here or click to uploadPNG, JPEG, WebP, BMP, GIF — max 4096px
Upload an image and click Remove Background. The MODNet AI model (~25MB) downloads on first use and is cached by your browser. WebGPU: ~1s, WASM: ~5s.
📷Drop image here or click to uploadPortrait photos work best for background blur
15px
Upload a portrait photo and click Apply Blur. Mask-based method runs background removal first (slower, more accurate). Simple method uses luminance approximation (faster).
📷Drop image here or click to uploadUpload photo with clear subject for best results
Sky
Beach
Office
Forest
Sunset
Night
Studio
Nature
1px
Upload an image, choose a new background, and click Replace Background. Runs background removal first to extract the subject, then composites onto your selected background.
📷Drop image here or click to uploadImage containing one or more faces
8
Upload a photo and click Detect & Blur Faces. Face detection runs locally using skin-color + edge heuristics. Detected faces are highlighted with blue boxes and processed.
📷Drop image here or click to uploadLow-resolution images benefit most from upscaling
80%
Upload a low-resolution image and choose your upscaling strategy. Fast mode uses bicubic interpolation + Unsharp Mask. AI mode runs real neural super-resolution locally.
Remove BG: Uses human matting AI model running locally. Optimized for portraits and subjects with clean edges. Your image never leaves your device.
Blur BG: Mask-based uses AI to isolate subject then blurs only the background. Simple method estimates depth from pixel luminance and applies graduated blur. All local processing.
Replace BG: Uses AI to extract the subject, then composites onto solid color, gradient, built-in templates (sky/beach/office/forest/sunset/night/studio/nature), or your own uploaded background. All local.
Face Blur: Skin-color region detection with aspect ratio filtering identifies face regions. Supports Gaussian blur, mosaic pixelation, or solid block cover. Works with multiple faces simultaneously. No external model download needed.
Super Res: Strategy A: Canvas bicubic interpolation + Unsharp Mask sharpening (instant, no download). Strategy B: AI (real neural super-resolution) via WebGPU/WASM. ~5-30MB on first download (cached afterwards); 2x/4x scales run purely through AI, 3x uses AI 2x then bicubic. WebGPU is faster, WASM works on any browser. Both run locally in your browser.
