Restore old photos with AI

A faded print or scratched family portrait can be hard to share. Photos Repair uses AI to create a restored version of your digital image, with optional colorization for black-and-white pictures. Keep your original as a reference: restoration is an interpretation of the image, not a recovery of the original scene.

Select my photos

100 credits each month for €9.99/month or €49.99/year. Extra packs require an active subscription.

Restored
Restored
Before
Before
Example result. Your result will depend on the original photo.

How it works

Start with a scan or a well-lit photograph of the print. Keep the camera parallel to the picture, avoid glare, and crop out the frame. A clear source gives the model more detail to work with. You can select JPEG, PNG, WebP or HEIC files, up to 20 MB each, and prepare up to 20 photos together.

  1. Select your photos and choose Restoration in the editor. Enable colorization if you want an interpretation in color.
  2. Choose 2K or 4K, sign in and check the credit cost. Buy a credit pack if your balance is insufficient; payment comes before processing.
  3. Start restoration, compare the result with the original and download your photos.

What to expect

AI may reduce visible scratches and fading, but severely damaged areas can produce invented details. Check faces, lettering and clothing carefully. Colorization suggests plausible colors; it cannot establish historical accuracy. Keep the original file and decide whether the result preserves what matters to you.

Choose 2K for a smaller output or 4K when you need more pixels. Print size also depends on the aspect ratio and the quality of the source. A larger output does not guarantee that lost detail has been recovered.

A documented test on an archive photo

We used a public-domain Library of Congress scan with the same image preparation and AI engine as this tool. The source was not artificially damaged. The app removes metadata and prepares a JPEG, then generates a 2K image with Nano Banana 2. The web example is encoded as JPEG.

Restored
Restored
Before
Before
Edwin Rosskam · 1941 · Library of Congress — Original archive scan
Source dimensions
735 × 1024 px
Generated dimensions
1792 × 2400 px
Setting
Restoration · 2K
Test date
2026-10-05

For Edwin Rosskam’s 1941 photograph of a mother and child, we kept colorization off. The result stays black and white and retains the scan’s film border. The faces remain soft despite the larger output: more pixels do not establish that fine facial detail has been recovered.

What to check before keeping the result

  • Compare eyes, mouths and expressions with the original at the same display size.
  • Check text, hands and clothing patterns for invented details.
  • Keep the scan and label the edited copy, particularly in a family archive.

This is one example for this setting, not a success-rate study or a comparison with another service. Your source and each generation can produce different results. Archive subjects are not customers or endorsers.

Source and generation record

Clear prices, before processing

One photo costs 5 credits in 2K or 10 credits in 4K. Both resolutions are available with every plan.

100 credits each month for €9.99/month or €49.99/year. Extra packs require an active subscription.

See prices

See the difference

Real photographs. Real results from our tool.

Explore three historical photographs processed by Photos Repair in 2K. Slide to compare. AI reconstructs details and estimates colors: always compare faces with the original.

Restored
Restored
Before
Before

A family portrait, colorized

Dorothea Lange · 1939

View the original archive
Restored
Restored
Before
Before

A faded scan, restored

Edwin Rosskam · 1941

View the original archive
Restored
After (HD)
Before
Before

A street scene, enhanced

Risdon Tillery · 1944

View the original archive

Public-domain archives: Library of Congress, Prints & Photographs Division, FSA/OWI Black-and-White Negatives. AI treatments by Photos Repair.

Choose the right tool for your photo

Restore old photos with AI | Photos Repair