Restore old photos in your browser with a full-image local model for fading, compression damage, blur, and low-resolution detail. Optionally colorize, enhance, and upscale 2x or 4x without uploading your photos to a business server. Missing pixels, colors, faces, and text may be inferred, so review important details before export.
Restore old photos in your browser with a full-image local model for fading, compression damage, blur, and low-resolution detail. Optionally colorize, enhance, and upscale 2x or 4x without uploading your photos to a business server. Missing pixels, colors, faces, and text may be inferred, so review important details before export.
Most processing is designed to run in the browser whenever possible, which makes the tool useful for quick personal or work files. For sensitive, production, or regulated material, test with a sample first and confirm the target platform's format, size, and quality requirements.
No. Reading, processing, comparison, export, and history storage happen in your browser and device storage. The tool does not upload source photos or results to the business backend.
It is better suited to fading, compression damage, light blur, and low-resolution scans. Long scratches, tears, large missing areas, and complex backgrounds are not guaranteed to be repaired automatically or recover real pixels.
No. Colorization is a tunable algorithmic preview. It does not prove the historical colors of clothing, skin, skies, or objects from a particular period.
Those areas may not contain enough original information, so the algorithm can only infer plausible-looking details. Review important faces, text, and evidence materials carefully.
Usually yes: repair damage first, colorize when needed, and apply conservative enhancement last. The workspace recalculates from the original and lets you disable any step.
PNG and JPG are available. Prefer PNG when you want to avoid another lossy compression step; JPG is useful for smaller sharing copies.