
Surface damage
Scratches, creases, fold lines, dust and water stains are each dealt with as their own problem. Damaged areas are rebuilt from the surrounding image rather than blurred over until they stop showing.


Upload a damaged photo and get it back as it was kept. Scratches, creases, stains and fading are removed one at a time, film grain is preserved, and faces come through unchanged.
AI old photo restoration repairs the damage a print collects over decades and rebuilds the detail underneath. This one works defect by defect rather than smoothing the whole frame, which is why the grain survives and the faces still look like the people in them.

One required scan, one mode, one optional line of notes. The work happens in what gets treated separately.

Scratches, creases, fold lines, dust and water stains are each dealt with as their own problem. Damaged areas are rebuilt from the surrounding image rather than blurred over until they stop showing.

A print that has lost its blacks gets them back. Tonal depth is recovered from what survived, and the result is the photograph as it was printed rather than a modern high-contrast version of it.

Film grain and historical texture are kept on purpose. An over-denoised face reads as a fake rather than as a restoration, so features are sharpened back rather than smoothed away and reinvented.
A flatbed scan gives the most to work with. A straight-on phone photo in even daylight is enough to start, as long as there is no glare across the surface.
The default repairs the damage and leaves a black-and-white photo black and white. Colour is a deliberate choice rather than something that happens to you.
Where the tear runs, who is in the frame, roughly what year, what colour the dress was. Notes are how a restoration stops guessing.
The restored photo lands on a fresh Virse canvas beside your scan. Re-run with different notes, run a colorized version alongside the repaired one, or push it further without exporting.
People holding photographs that only exist once.

Work through a box of prints in one place, with every original still sitting beside what came back from it.

Recover faces and details clearly enough to identify relatives across generations of the same album.

Bring one photograph to print quality for the day it matters, without sending the original away.






The tempting shortcut is to smooth everything until the damage stops showing. That is also how a restoration starts looking fake.
One-shot restorers
The Virse restorer
Upload it here and the restoration opens on a fresh Virse canvas you can keep working on, with nothing running until you press Start.