Photo Restoration Prompt For ChatGPT (Free Copy-Paste)

Photo Restoration Prompt for ChatGPT: Fix Torn, Faded, and Blurry Photos

Introduction

I have tested dozens of photo restoration prompts on ChatGPT, and the fastest way to fix an old or damaged photo is to upload a clear picture and tell the AI exactly what is wrong with it, torn edges, faded color, scratches, or blur, instead of just saying “make this look better.” A vague prompt like that almost always gives you a soft, plastic-looking face, because the AI ends up smoothing the damage and the person’s features at the same time.

The real trick with old photo restoration is picking the right prompt for the right problem. A ripped photo needs a very different set of instructions than a yellowed one, and a photo you clicked on your phone from a table needs a completely different approach if you want it to look like a real digital scan. Below are six simple, tested prompts for the most common photo problems people run into with old family pictures, from torn corners to a phone photo that needs to look like it came off a proper scanner. Each one works directly in ChatGPT, and the same prompt also works if you are using Gemini AI instead.

6 Ready to Use Photo Restoration Prompt

Gemini AI Prompt

Torn Old Photo Repair Prompt

Repair this torn photograph by fixing all rips, cuts, and missing corners. Reconstruct the missing parts using the surrounding image details so it looks complete and undamaged. Keep the faces, clothes, and background exactly the same as the original. Do not add new people or objects. Keep it black and white if the original is black and white, or keep original colors if it is a color photo. Final result should look like a clean, complete photograph with no visible tear or missing area.

Gemini AI Prompt

Faded Photo Color Fix Prompt

Fix this faded and yellowed photograph by restoring its natural, accurate colors. Remove the yellow or orange tint and bring back real skin tones, clothing colors, and background colors. Keep the same people, poses, and expressions exactly as in the original. Do not change anything else in the photo. Final result should look fresh and naturally colored, like the photo was taken recently, not artificial or overly bright.

Gemini AI Prompt

Scratched Photo Cleanup Prompt

Remove all scratches, lines, and dust spots from this photograph. Clean the surface completely while keeping every face, object, and background detail exactly the same as the original. Do not smooth or blur the faces — keep them sharp and natural. Do not change the photo’s colors or lighting. Final result should be a clean, scratch-free version of the exact same photo.

Gemini AI Prompt

Blurry Old Photo Sharpening Prompt

Sharpen this blurry photograph and improve its overall clarity. Enhance the details in the face, eyes, hair, and clothing without changing anyone’s identity or facial features. Keep the same expressions, pose, and background as the original. Do not add new details that are not visible in the blurry version. Final result should look clear, crisp, and naturally sharp, like it was taken with a good camera.

Gemini AI Prompt

Black and White to Color Photo Prompt

Convert this black and white photograph into a realistic colored photo. Add natural skin tones, realistic clothing colors, and accurate background colors based on the time period and setting of the photo. Keep every face, pose, and expression exactly the same as the original. Do not make the colors too bright or artificial — keep them natural and true to life. Final result should look like a real color photograph from that era, not a cartoon or overly saturated image.

Gemini AI Prompt

Phone-Captured Photo to Digital HD Scan

Convert this phone-captured image of a physical photograph into a clean, high-definition digital scan. Remove all background elements such as the table, hands, wall, or any surface the photo was placed on, and extract only the photograph itself with straight, clean edges. Eliminate glare, reflections, shadows, and uneven lighting caused by the camera flash or angle. Correct any perspective distortion so the photo appears perfectly flat and rectangular, as if scanned directly on a flatscanner. Remove visible curling, bending, or physical creases of the printed photo paper. Enhance sharpness and resolution to high definition quality without altering the identity, facial features, expressions, or clothing of the people in the photo. Restore natural, accurate colors if the original has faded or has a color cast. Do not add any new elements, people, or objects. The final result should look like a professionally scanned, digitally archived version of the original photograph — flat, sharp, color-accurate, and free of any camera artifacts or background.

How to use

Always start with the clearest photo you have, since a blurry or dark scan gives ChatGPT less to work with. For the torn photo repair prompt, make sure the parts of the photo that are not damaged are clearly visible, because ChatGPT looks at those good parts to figure out what the torn or missing area should look like. If you only have a phone photo of a printed picture lying on a table, use the phone-to-digital-scan prompt first, before trying anything else on that image, since it clears out the background, glare, and tilted angle so the rest of the prompts have a clean photo to work on. For the black-and-white to color prompt, I always mention the rough decade or place the photo was taken in, since that small detail helps ChatGPT choose skin tones and clothing colors that actually match the period instead of guessing something too modern.

Tips for best results

In my own testing, the scratched photo cleanup prompt works much better on a higher-quality scan, because on a low-quality image, small scratches can get mixed up with hair or fabric texture and end up staying in the final result. With the blurry photo sharpening prompt, I got the most reliable results when the face was at least somewhat visible in the original photo. On photos that are completely out of focus, ChatGPT has to guess too much, and the face can come out looking slightly off. One more thing I noticed with the phone-to-digital-scan prompt: if there is strong flash glare right across someone’s face, it helps to retake the phone photo from a slight angle first, since glare over a face is much harder for the AI to clean up than glare on the background.

FAQs

Can I restore a photo I only have as a phone picture, not a scanner scan?

Yes, that’s exactly what the phone-captured photo prompt is for. It removes the table, hands, or wall in the background, straightens out the angle so the photo looks flat instead of tilted, and cleans up glare or shadows from your camera flash, so you end up with a photo that looks like it was scanned properly, even though you just clicked it with your phone.

Will the black-and-white to color conversion pick accurate colors for an old photo?

ChatGPT can’t know the exact original colors since a black-and-white photo doesn’t hold that information, so it makes a smart guess based on realistic skin tones and the kind of clothing and settings common to that period. If you tell it roughly when or where the photo was taken, the AI colorization usually looks a lot more believable.

Does the scratch removal prompt risk blurring the person’s face?

No, this prompt is written so ChatGPT keeps the face sharp while it cleans the scratches off the surface, since removing scratches and softening a face are two separate things. If a result still comes out looking a bit soft, just run the same photo restoration prompt again on a slightly higher-quality version of the photo and it usually fixes itself.

Conclusion

Old family photos usually have some kind of damage, a tear, a faded corner, a scratch, or just years of sitting in a drawer, and that’s completely normal. What matters is that you don’t need any editing skills or expensive software to fix them anymore. A good photo restoration prompt on ChatGPT can repair tears, bring back faded colors, clean up scratches, sharpen blur, add color to a black-and-white photo, or turn a rushed phone picture into a clean digital scan, all from a simple text prompt. Try the one that matches your photo’s exact problem, and if the first result isn’t perfect, running it again on a slightly better quality image usually gets you there.

Priya Sharma

Hi, I’m Priya Sharma, the creator of ImagePromptLab.

I research and test AI photo prompts across Google Gemini, and ChatGPT before publishing anything on this site. My focus is on South Asian portraits, desi styles, cultural occasions, modest fashion looks and everything in between that most AI tools don’t get right. Every prompt here is personally tested and refined to make sure it actually delivers good results. My goal is simple, to give anyone access to free, ready to use prompts that create beautiful, professional looking AI photos without any technical skills required.