How to restore old photos online, for free

No software to install, no account, no watermark.

One photo at a time here. Need a batch, a different mode, or the before/after slider? Open the full tool.

Looking for photos you deleted?

This page improves the quality of a photo you still have; it cannot bring back a file that was deleted. For that, look in your phone's recently deleted folder first, then in your iCloud or Google Photos backup. Once the file is back, you can restore it here.

Why photos fade, and how restoring helps

Old photographs lose sharpness for all kinds of reasons: prints yellow and fade with age, scans soften fine detail, and early digital cameras simply couldn't capture much resolution. PicReviver uses an advanced image model trained on millions of photos to rebuild the detail that's been lost: sharpening edges, redrawing texture in faces and clothing, and cleaning up the soft, washed-out look of an ageing picture, without making it look over-processed.

Scratches, cracks and creases

A print that's been handled for decades picks up a map of its own history: fine surface scratches, white crease lines where it was once folded, hairline cracks in the emulsion. The tool treats these the way it treats any lost detail: it reads the intact area around each mark and rebuilds what the surface underneath most likely looked like. Thin scratches and light creases usually soften away almost completely, because there's plenty of surrounding detail to work from. A deep crease running straight through a face is harder: there's less to go on exactly where the damage is worst, so you'll see a real improvement rather than a flawless repair. The one thing no tool can do is invent a piece that's physically gone. If a corner has been torn away or a chunk is missing, there's no detail left there to rebuild from, and restoration only works with what's still in the photo.

Water damage, stains and heavy fading

Water is one of the most common ways old photos get ruined: a leaky box in the attic, a flood, a spilt drink. It leaves tide-lines, cloudy blooms and brown or yellow staining, and it often lifts the colour unevenly, so one part of a picture looks bleached while another goes dark. Fading works slowly in the same direction: sunlight and time wash the contrast out until faces go flat and grey. Restoration tackles both by rebuilding the contrast and texture the damage flattened, pulling detail back out of the murky areas; it does not correct discolouration, so a stain keeps its colour. Heavy staining that has eaten right through the image won't vanish entirely, but a faded, water-marked print almost always comes back looking markedly clearer and closer to the photo it used to be. If it's black-and-white and you'd like it in colour afterwards, run the restored copy through colourising.

How to restore a photo

Best for

• Faded family photos and old prints
• Soft, low-resolution scans of vintage pictures
• Blurry snapshots from older phones and cameras (or see sharpening a blurry photo on its own)
• Small images you want larger and sharper (also try upscaling)
• Black & white photos you'd like in colour (try colourising)

What restoration can't do (being honest)

We'd rather set expectations than oversell. Restoration rebuilds detail that's faint, soft or damaged; it can't recreate detail that was never captured or has been completely destroyed. A tiny, extremely low-resolution scan has a ceiling on how much can come back. A face that's a soft grey smudge with no features left gives the tool almost nothing to reconstruct from. Physically missing pieces can't be filled from nothing. What you can count on: a faded, scratched, soft or water-damaged photo comes back noticeably clearer, and when the tool can't fully fix something it errs towards a natural result rather than an obviously fake one. If the first result isn't quite right, a cleaner scan of the original print, made the way our scanning guide describes, often makes the difference.

What actually gets rejected, and what doesn't

Between 3 August and 2 October 2026, 5,436 photos were uploaded to PicReviver's restore, sharpen and colourise tools. The small tool on this page was used 48 times in that period; most people restore old photos with the main tool on the homepage, which ran 2,754 times. File size was rarely the problem: the average photo was 1.6 MB, and only one went over the 99 MB limit. Pixel count was the more common snag. 29 photos were refused for being between 45 and 103 megapixels, which usually means a print was scanned at a very high DPI. What people ran into most was the free daily limit of ten runs, reached 149 times. So if a scan of an old print looks huge on your computer, lowering the scan resolution is far more likely to help than shrinking the file.

Restore old photos free, from anywhere, with no software to buy

Is there a catch? No paywall, no software to buy and no account to create. You can restore old photos online free from any browser, on a phone or a desktop, because the work happens on our GPU and the finished file downloads straight from the tab. If you have tried older restoration tools that are really filters, this one is different: it rebuilds detail with a trained model instead of laying a preset over the top. To restore old photographs or restore old pictures, the steps are the ones described above: upload, wait a few seconds, download the full-size result. A batch of five works as well as a single restore old photo upload, and a print that is soft as well as faded gets both fixed in the same pass.

There is no shop to visit and no prompt to write: upload from wherever you are and the finished file comes back to the same device. If a fragile print needs physical repair, a local conservator is the right call. For the digital file, the free tier and a paid pass use the same trained model.

Combining restoration with colour, sharpening and upscaling

Restoring and colourising are two separate steps. Restore first to rebuild the detail, download the result, then colourise that copy if the print started in black and white. Restore keeps the colours it finds, so it will not fix a yellowed photo; colourising replaces the colour altogether with a prediction made from the grey tones. For a photo that is too small as well as faded, the restore pass already enlarges, and the upscaler can take a restored copy further. The same model handles a digital scan, a slide or a phone photo of a print, since it only looks at the pixels it is given.

Recovering a lost file versus repairing the one you have

Here you can photo recover online only in the quality sense: rebuilding detail that fading, scratches or creases wore away, not getting back a deleted file. To repair old photos, upload a scan or a digital copy of the print. An old damaged photo usually has several problems at once (fading, scratches, a crease or two), and one upload handles the whole mix within the limits described above. Picture restoration of this kind works the same for a print, a slide or a digital scan, and an old image with plenty of surviving detail responds best. A sharp, evenly lit scan is the best input for AI restoration.

The model was trained on millions of photo pairs that were degraded on purpose with blur, noise and compression, which is why it copes with a messy real scan. Most ai photo restoration tools bundle sharpening, colourising and repair under one name; this page runs the sharpening-and-repair half, and colourising is a separate step with its own model. What the AI is doing is drawing a best guess at the lost detail, learned from those training photos. That suits the kind of restoration old pictures need, but it is not a record of what was actually there.

Free access, and how this differs from a manual photo editor

The free photo restoration tier covers up to five photos per run and ten runs a day, at full resolution, with no watermark and no card, whether the photo is mildly faded or properly damaged. That is what old photo restoration online free means here. A paid pass exists for big batches, but photo restoration online free runs the same model at the same quality. If you are wondering how to restore old pictures without a specialist studio, it comes down to three steps: upload the print, let the model rebuild the detail, download the full-size file. A scanned print, a phone photo or an old digital copy all go through the same pass.

This is not a full photo editor with layers and brushes. A retoucher used to touch up a print with a pencil; this does the digital equivalent automatically, in seconds. It also doesn't turn an old photo into a new one: it rebuilds plausible detail from what survives rather than manufacturing a picture that never existed, and that is the real difference from a sharpening filter that only adds contrast. A soft scan or slight camera shake responds well; a badly out-of-focus original has a lower ceiling. Each photo restore takes a few seconds, and every photo repair pass runs on our own GPU hardware, with nothing to install.

If what you're restoring is a scanned document page rather than a photograph, a different tool fits better: ScanReviver's clean scanned document straightens it, crops the desk away and evens out the lighting. For photos, you can restore old photo online free of charge here, up to five at a time, with no card and no watermark.

Before you ask

By Sam Ridder — I build and run PicReviver on my own hardware, on my own. Who I am.