AI photo enhancement used to mean signing up for a SaaS, uploading your family photos to a stranger's GPU cluster, and paying per image after a free trial. The AI Image Tools hub collapses that workflow into a single browser tab: the same upscaling, restoration and colorisation models, but running on your own device via WebAssembly and (where available) WebGPU. Your photos never leave the page, and there is no per-image charge.
The flagship tool is Super Resolution: choose 2x or 4x and the model reconstructs detail that bicubic upscaling cannot recover, particularly on faces, text and natural textures. Restore Photo targets old scanned prints — removing scratches, fixing fading, denoising film grain — and is the go-to tool for digitising a parent's photo album. Colorise Photo applies tasteful colour to black-and-white pictures, with a result that mimics period-appropriate palettes rather than the over-saturated look of older auto-colourisers.
For more surgical fixes, Unblur Image sharpens motion-blurred shots from phone cameras, Remove Background isolates subjects cleanly enough for e-commerce listings, Remove Objects and Remove Watermark erase unwanted elements with brush-based masks, and Cleanup Picture handles the finer blemishes and dust spots. All of these models run on-device; the first use downloads the model weights (typically 20–80 MB) and caches them so later runs are instant.
Performance varies with hardware: a recent MacBook with WebGPU finishes a 4x upscale on a 12 MP photo in roughly three to eight seconds, while a five-year-old laptop on the WASM-only fallback may take 30–60 seconds. Both produce the same output. Memory usage is the practical ceiling — beyond about 20 megapixels you may need to downscale first or use a desktop tool. Everything else (no quota, no watermark, no signup) is designed to remove the friction that has made AI enhancement feel like a paid niche product.
This hub pairs naturally with Image Editor for cropping the result, and with Image Converter for format swaps before sharing. If you have decades of scanned family photos waiting in a drawer, this collection is the most cost-effective way to digitise and restore them — start with Restore Photo for damaged scans, Super Resolution for low-resolution sources, and Colorise Photo for archival black-and-white prints.