Restore Old Photos
Run the whole clean-up on an old photo in one pass: reduce grain, remove fine scratches and dust, and bring back faded contrast and colour - with each stage under your control.
- Denoise, scratches and fading together
- Choose each stage
- Before / after with zoom
Classic image processing Grain, scratches and fading are corrected with classic image-processing methods. If an AI restoration stage is used for a photo, the result says so and labels that detail as estimated. The optional 2x enlargement uses a small learned model that sharpens; it does not invent detail.
Choose a photo
Before
After
Restore Old Photo workspace
Drop your photo here
or choose a file · on a phone you can take a photo or pick one from your gallery
Ready
Working
Preparing your photo…
- Uploading
- Waiting in line
- Processing
- Loading the result
Most photos take under a minute. Larger photos and busy times take longer.
Finished
Compare the result
Drag the divider or use the arrow keys to compare. Zoom in to check fine detail; when zoomed, drag the picture or use the arrow keys on it to move around.
That did not work
What to expect
- Works best on scans and phone photos of prints with fine scratches, dust and fading.
- Large tears, missing corners and big stains need Repair Damaged Photo, where you mark the area.
- Scratch removal fills thin lines with estimates from nearby pixels.
Your photo is processed on a2z.tools servers. No third-party AI provider receives it, it is not used for training, and the upload and result are deleted automatically 60 minutes after processing finishes.
Related Tools
More tools in the same category
Other Tools
Popular tools from across A2Z
What Restore Old Photo does
Old prints usually have several problems at once: grain from the film or the scanner, thin scratches and dust, and a washed-out, yellowed tone. Restore Old Photo tackles those three together, so you can get a cleaner copy of a family photo without running three separate tools.
What happens to the photo
The photo is re-encoded on our server without its metadata, then passed through the stages you left switched on. Noise reduction uses non-local means filtering, which averages similar patches of the picture rather than blurring everything. Scratch removal first detects thin bright or dark lines and dust spots, then fills them by inpainting, a classic method that continues the surrounding pixels into each mark. Fade correction stretches the tones and balances colour casts using contrast equalisation.
These standard stages are not generative AI. They work with what is already in the photo, which is why they are predictable, and also why they cannot recreate detail that was never captured. If an AI restoration stage is used for a photo, the result says so and labels that detail as estimated, and fade correction is then skipped because that stage already adjusts tone and colour. If you switch on the 2x enlargement, a small learned super-resolution model enlarges the result; it keeps edges a little sharper than ordinary resizing but does not add detail.
Getting a better result
A sharp, evenly lit scan matters more than any setting. Crop away the album page and frame before uploading, and try the default settings first. If real detail such as hair is being removed, lower the scratch sensitivity; if the photo looks waxy, choose a lighter noise level.
When to use it
- A scanned family print with dust specks, fine scratches and a yellowed, low-contrast look.
- A phone photo of an old album page that looks grainy and flat.
- A quick first pass before deciding whether a photo needs the more specialised repair tools.
How to use it
- Scan the print at 300 DPI or photograph it flat in even daylight, then open Restore Old Photo.
- Drop the photo on the page or choose it from your files or phone gallery.
- Keep or switch off each stage - noise reduction, scratch removal and fade correction - and set their strength.
- Choose Process photo and wait while the job is queued and processed.
- Drag the before and after divider, zoom to 2x or 4x to check detail, then download as PNG, JPG or WebP.
Limitations
- Tears, holes, large stains and missing corners are not filled automatically; mark them in Repair Damaged Photo.
- Automatic scratch detection can miss faint lines or mistake fine hair and fabric texture for scratches.
- Colours that have faded away completely cannot be brought back, and black and white photos stay black and white.
- Strong noise reduction softens fine texture together with the grain.
Privacy
Your photo is processed on a2z.tools servers. It is not sent to any third-party AI provider and is not used for training. The upload and the result are deleted automatically about 60 minutes after processing finishes, and straight away if processing fails or is cancelled. Results are returned only to the browser session that uploaded the photo.
Rate this tool
Was this tool useful? Your feedback helps us improve it.
Key facts
- Runs noise reduction, scratch and dust removal and fade correction in a single job.
- Every stage can be switched off, with its own strength or sensitivity setting.
- Uses classic image-processing methods; any AI-estimated stage is named on the result, and filled scratches are estimates from nearby pixels.
- Optional 2x enlargement with a small learned model (FSRCNN) that sharpens but does not invent detail.
- Accepts JPG, PNG, WebP, BMP, TIFF and HEIC photos within the size limits shown on the page.
- Results can be downloaded as PNG, JPG or WebP.
How to use it
- Scan the print at 300 DPI or photograph it flat in even daylight, then open Restore Old Photo.
- Drop the photo on the page or choose it from your files or phone gallery.
- Keep or switch off each stage - noise reduction, scratch removal and fade correction - and set their strength.
- Choose Process photo and wait while the job is queued and processed.
- Drag the before and after divider, zoom to 2x or 4x to check detail, then download as PNG, JPG or WebP.
Frequently asked questions
What does Restore Old Photo change in my picture?
It reduces grain and noise, fills fine scratches and dust spots it detects, and restores contrast and colour balance in a faded print. You can switch each of those stages off, and nothing else in the photo is altered.
Is this generative AI that redraws the photo?
The standard stages are classic image processing: filtering, inpainting and contrast equalisation, which only rearrange information already in the photo. If an AI restoration or face stage is used for a photo, the result says so and labels that detail as an estimate. The optional 2x enlargement uses a small learned model that sharpens edges but does not invent new detail.
Why were some scratches left in the photo?
Automatic detection looks for thin lines and small spots that stand out from their surroundings. Faint scratches, or ones that run along real edges, can be missed. Raise the scratch sensitivity, or use Remove Scratches from Photo, where you can correct the detected mask with a brush.
Can it fix a torn or badly damaged photo?
Not automatically. Large tears, holes, stains and missing corners need you to mark the damage. Use Repair Damaged Photo for those areas, then continue with the rest of the workflow.
Will it add colour to a black and white photo?
No. Fade correction strengthens colour that is still present in a colour print, and a black and white photo stays black and white. To add colour, restore the photo first, download it, and then use the separate Photo Colorizer, whose colours are predictions.
Which settings should I start with?
Start with the defaults: medium noise reduction, scratch removal at medium sensitivity and fade correction at 80 percent. Compare the result with the zoom, then lower the scratch sensitivity if real detail disappears, or turn on Keep the sepia tone if you prefer the original warm look.
Last reviewed: