AI Photo Restoration

Restore old photos

Add your old photo

One clear source photo

Use a clear scan or a straight-on photo of the print. Keep the original file.

JPEG, PNG, or WebP

Or try an example

Choose a restoration direction6 repair or color goals

Left: original. Right: illustrative result. Presets aim to keep film grain; use Custom for a sepia finish.

Resolution

Aspect Ratio

Output

Your generated results will appear below.

My Creations

Give an old photo a clearer finish Choose a repair direction and compare the result carefully. Examples are illustrations; generated details are not recovered historical facts.

Illustrative before / after

HOW IT WORKS

From your photo to one edited image.

Follow the illustrated workflow from upload to review. The shown looks are examples, not guaranteed generations.

Illustrative old portrait before restoration

ILLUSTRATIVE OLD PHOTO

STEP 01 · Upload

Start with the best scan

Upload a clear scan; keep the untouched original.

Upload a photo

Damage repair

Revive faded colors

Restore & colorize

EXAMPLE DIRECTIONS

STEP 02 · Direct

Match the photo's problem

Repair damage, revive faded colors, or colorize a monochrome photo.

Choose a direction

SAME SOURCE PHOTO

ASPECT RATIO 1:1 4:5 16:9 RESOLUTION 1K 2K 4K

CHOOSE YOUR FRAME

STEP 03 · Generate

Choose output settings

Set the frame and resolution; check the credit cost.

Set the frame

Illustrative edited result from the fictional source

ILLUSTRATIVE LOOK

STEP 04 · Review

Inspect faces and texture

Compare details with the original before saving.

Start editing

MADE FOR EVERYDAY CREATORS

Photo-restoration examples for three common problems.

Match your photo to three repair goals: light scratches, faded color, or black-and-white colorization. Each source/result pair is a fictional illustration. Keep your untouched scan and compare faces, texture, and tones before saving.

Repair scratches in an old black-and-white photo

Choose Damage repair for light scratches, small stains, and thin fold marks on an old print. Upload a clear scan and compare the eyes, mouth, clothing, and film grain with the untouched original. AI photo restoration can reduce visible damage, but large tears and missing facial detail may require manual work.

Repair light scratches

Fictional black-and-white portrait with light scratches and surface stains

Illustrative black-and-white portrait with light scratches cleaned and original tones retained

Restore faded colors in a family snapshot

Choose Revive faded colors for an originally colored photo with washed-out tones or an unwanted cast. The direction uses surviving colors rather than adding color to a monochrome image. Check skin tones and known clothing colors against the print. Colors that have disappeared completely cannot be verified from the damaged photo.

Revive a faded color photo

Fictional originally colored family snapshot with washed-out colors and a magenta-yellow cast

Illustrative color family snapshot with fading and a magenta-yellow cast reduced, retaining the same people and clothing

Colorize a black-and-white family photo

Choose Restore & colorize to explore a color version while reducing light damage. Keep the original scan beside the result and inspect every face. In Custom, describe colors you know, such as the blue dress or brown coat. AI colorization supplies plausible colors; it does not recover verified historical colors.

Explore photo colorization

Fictional black-and-white family garden photo before repair and colorization

Illustrative black-and-white family garden photo restored and colorized in muted natural colors

FAQ

AI Photo Restoration: common questions.

How can I restore an old photo without changing the face?

Natural restore asks for a gentle repair while retaining facial structure, expression, and film grain. Upload your clearest scan; in Custom, name distinctive features to keep. Compare the eyes, mouth, and hairline with the original. AI can still alter identity, and there is no face lock. A sharper generated face is not proof of a recovered likeness.

Can AI restore a blurry old photo without making it look artificial?

Portrait clarity requests modest improvement to visible facial, hair, and clothing edges while retaining grain and natural texture. Use the original scan rather than a compressed screenshot. Strong blur may leave insufficient detail, so AI can invent a sharper-looking face. A 2K or 4K output does not recover verified information missing from the source.

Can I repair scratches, stains, and fold marks in an old photo?

Damage repair targets light scratches, small stains, and thin fold lines while asking to retain faces, clothing, tones, and grain. Keep an untouched scan for comparison. Large tears, missing faces, and heavily obscured areas can produce guesses. Consider manual restoration when accurate reconstruction matters; this AI workflow cannot guarantee repair of every defect.

Can I restore faded photo colors without adding new colors?

Revive faded colors is designed for an originally colored print. It requests reduced fading and color casts using surviving tones, rather than colorizing a monochrome photo. In Custom, name known colors, such as a blue dress. Compare with the original print or other evidence; AI cannot verify pigments or colors that have completely disappeared.

Can I restore a black-and-white photo without colorizing it?

Natural restore keeps the source's monochrome, sepia, or color treatment. Choose Neutral B&W for restrained gray tones, or use Custom to retain a specific sepia finish. Restore & colorize deliberately adds plausible colors to a separate interpretation. Keep the untouched original; generated colors and details do not establish historical accuracy.

Should I scan an old photo or upload a phone picture of the print?

A clear scan provides more visible detail for photo restoration. A straight-on phone picture can also be used if the whole print is visible with even light and no glare. Upload JPEG, PNG, or WebP and retain the original file. The editor does not provide a dedicated scanning, perspective-correction, or glare-removal control.

What prompt should I use to restore an old family photo?

Describe the damage and what should stay. For example: Reduce light scratches and fading; keep the photo black and white, with the same faces, expressions, clothing, background, and film grain. Custom accepts up to 1,000 characters or one tone reference, used separately. The reference guides finish, not missing facial identity or factual color recovery.

Can I restore old photos online for free?

You can explore the restoration editor without an account; generation requires sign-in and credits. New accounts receive 40 credits for 30 days. Check the displayed cost before submitting. Failed generation returns credits, while a successfully generated image you dislike still uses them. Another repair attempt is a separate request, so restoration is not unlimited or guaranteed usable.

START WITH YOUR PHOTO

Keep the memory. Review every detail.

Choose a source image, give the model one clear direction, and review the result before using it.