The short answer: upload old photos (up to 50 at once), and AI removes cracks, creases, stains and scratches, restores believable tonality and facial detail; grain and period character stay, and facial features are never invented. The restored photo is still a photo of the same person. Want color next? Send it to the film colorizer. Want it bigger and sharper? Send it to the upscaler.
What old-photo restoration does
| It does | It doesn't |
|---|---|
| Remove cracks, creases, stains, scratches and mildew | Invent large missing areas |
| Restore believable tonality and facial detail | Redraw or swap facial features |
| Keep grain, tone and period character | Produce a plastic, over-smoothed look |
| Support color and black-and-white | Colorize automatically (that's the colorizer's job) |
The boundary is deliberate: restoration removes damage; it doesn't repaint the person. Old family photos matter because they show real people; the tool repairs the wear of time, not the face.
What it fixes, and how to prepare
Best results: small cracks, creases, stains, mildew spots, yellowing, fading and slight blur.
Never invented: large missing areas, big mildew patches or missing facial features; you can try, but test one image first.
Preparing scans and captures:
- Crop black borders off scans;
- Photograph straight-on with a phone, flash off, glare minimized;
- The higher the resolution, the better the result.
How to restore step by step
- Upload old photos: one or many, up to 50; JPG / PNG / WebP, each up to 10 MB; the batch shares one restoration standard, exactly right for albums and archive digitization;
- Optional: a note, e.g. "keep the grain, focus on the crease at bottom right";
- Download: compare before/after and save; files keep their names under the
eyesme-restoredprefix.
Four scenes in practice
A crease across the face: the hardest restoration, since the features must stay exactly as they are while the damage disappears.
Family group: a 1960s living-room portrait loses its fading and creases, and every face stays the same face:
All four faces keep their identity while color and tonality recover
- Sepia portrait: emulsion cracks and small stains cleaned, skin texture kept:
Cracks and spots removed; facial light and shadow readable again
- Streets and scans: a 1950s street scene's dust and scratches, a garden couple's mid-frame crease, both rebuilt from surrounding texture:
Building and street detail recover while the period grain stays
The mid-frame crease disappears; fading and dust cleaned in one pass
After restoration
Restoration is step one; the next steps have their own tools:
- Add color: send black-and-white photos to AI film colorization with its 12 film looks, leaving people and tonal structure untouched;
- Go bigger: send the restored photo to AI upscaling, which doubles width and height with rebuilt texture;
- Suggested flow: restore → colorize → upscale; each step's output feeds the next.
vs manual retouching and general AI editors
| Manual retouching / old-photo apps | AI old-photo restoration | |
|---|---|---|
| Workflow | Clone stamp and repair tools, spot by spot | Upload and restore in one click |
| Cracks & stains | Cleaned by hand, mark by mark | Removed automatically |
| Tonality & faces | Easy to over-brighten or lose detail | Believable tonality and facial detail restored |
| Period character | Can look plastic | Grain, color and period feel kept |
| Batch | One image at a time | Up to 50 images, one standard |
| General AI editors | AI old-photo restoration | |
|---|---|---|
| Getting started | Rewrite prompts and tune parameters | Upload and restore; add notes if needed |
| Feature fidelity | Can redraw or swap faces | Original features, expression and face shape kept |
| Period character | Often over-smooths | Grain and period color preserved |
| Batch consistency | Tune each image; style drifts | One restoration standard for the batch |
In one line: manual retouching wins on per-pixel control, general AI on freedom, and restoration wins on "repair the damage, not the person, period feel intact, one standard per batch".
Frequently asked questions
How does photo restoration work?
Upload an old photo and click start; AI removes cracks, creases, stains and scratches and restores believable tonality and facial detail while keeping the original grain and period character.
Which problems can it fix?
Small cracks, creases, stains, mildew spots, yellowing, fading and slight blur work best; severe damage, large missing areas or missing features are not invented.
Will it change the person's face?
It preserves the original features, expression, face shape and period character as much as possible and only repairs damage and tonality.
Does it support color and black-and-white? Does it colorize?
Both are supported; the tool only repairs and does not colorize automatically; use the film colorizer or add a note.
Do scans and phone captures work?
Yes. Higher resolution and less glare work better; crop black borders from scans and shoot straight-on without flash.
Can I process a batch?
Yes, up to 50 images at once with the same restoration standard; ideal for family albums and archive digitization.
What can I do after restoring?
Colorize (film colorizer) or upscale (photo upscaler); the suggested order is restore → colorize → upscale.
Which formats and output sizes are supported?
JPG, PNG or WebP, up to 10 MB each; output size adapts to the input, and the result card shows actual dimensions.
Features and details reflect the tool as of November 12, 2026 and may evolve; the before/after examples are from the tool's real demo assets, with the cover composited from the damaged original and restored halves.
Ready to make old photos readable again? Eyesme's AI Photo Restoration takes up to 50 photos at once: cracks and stains gone, tonality restored, faces and period character intact.




