The Ultimate AI Prompt for Restoring Old & Damaged Photos

Blogger Post Template

Professional Archival Vintage Photo Restoration Prompt | AI Photo Prompt

Transform damaged, faded, and scratched historical photographs into pristine, high-resolution memories with this advanced restoration prompt.

Act as a master archival photo conservator with 30 years of experience in restoring antique, damaged, and faded photographs. Your task is to comprehensively restore the uploaded vintage/old photo. Follow these forensic-level instructions with absolute precision: Damage Repair: - Remove ALL scratches, dust specks, mold spots, creases, fold lines, tears, and water stains. - Seamlessly inpaint missing or torn sections (e.g., torn corners, missing edges, or cracked faces) using logical context from the surrounding environment. Do NOT invent new features—only reconstruct what is plausibly there. Tonal & Contrast Restoration: - Correct faded highlights and crushed shadows. Restore the proper dynamic range of the original image. - If the photo is black and white, ensure neutral grays with absolutely zero color cast (no sepia, blue, or green tints). - If the photo is faded sepia or vintage color, balance the color channels realistically and subtly. Do NOT oversaturate. Sharpening & Deblurring: - Gently reduce mild motion blur, camera shake, and focus softness. Sharpen edges naturally without creating harsh digital halos or white outlines. Skin & Texture Integrity: - Preserve every wrinkle, scar, pore, and freckle exactly as they are. Do NOT smooth, plasticize, or "beautify" the skin. - Do NOT change facial features, bone structure, eye shape, hair pattern, or body proportions. - Enhance fabric textures, wood grains, and environmental details naturally. Historical Authenticity: - Maintain the authentic vintage character of the photograph. Do NOT make it look like a modern digital RAW file. - Add an extremely subtle, organic film grain over the final output so it still feels like a classic photograph from its original era. Output Format: - Output the restored image in high resolution with the same aspect ratio as the original. - Preserve the original framing and crop—do NOT zoom in, zoom out, or rotate. Negative Constraints: - Absolutely avoid AI "hallucinations" (inventing teeth, fingers, facial hair, clothing patterns, or background objects that don't exist). - Avoid over-brightening, over-sharpening, or artificial HDR effects. - Avoid color bleeding across edges. - Avoid making the image look like a cartoon, oil painting, or plastic mannequin. - Avoid changing the subject's expression, age, or ethnicity.

☕ Support Us
Copy the prompt ➡️ Open your preferred AI image-generation tool, such as Midjourney, ChatGPT, Google Gemini, or Stable Diffusion ➡️ Upload your photo if required➡️ Paste the prompt into the tool ➡️ Generate your image and enjoy the result!

Introduction

An AI photo prompt is a descriptive text instruction given to Artificial Intelligence models like Midjourney, ChatGPT, or Google Gemini. It acts as a creative blueprint, telling the AI exactly what kind of image to generate, including details about the subject, lighting, environment, and artistic style. A well-crafted prompt is the secret to unlocking stunning, highly detailed AI-generated artwork.

What This Prompt Creates

This prompt functions as a virtual photo restorer. It takes an uploaded, damaged vintage photograph and acts as an expert archival conservator to repair physical damage like scratches, dust, creases, and tears. It restores tonal range, corrects fading, and sharpens blurriness, all while strictly maintaining historical authenticity by preserving original textures, facial features, and adding a subtle organic film grain.

The AI Photo Prompt

This prompt utilizes a master-persona definition ("master archival photo conservator") combined with precise, forensic-level technical instructions. Keywords like "inpaint," "neutral grays," and "dynamic range" guide the technical restoration, while a heavy emphasis on preserving "skin & texture integrity" and strong "negative constraints" prevents common AI mistakes like smoothing faces or hallucinating new details.

Recommended Settings

  • Best Aspect Ratio: Maintain original aspect ratio of the input photo.
  • Lighting & Camera Style: High-resolution flatbed scan style; neutral, archival lighting.

How to Use This Prompt

This prompt is designed for AI tools that allow image uploading and editing (Image-to-Image). Upload your scanned old photo to ChatGPT, Google Gemini, or use an image-to-image workflow in Midjourney/Stable Diffusion. Paste the provided text prompt alongside the image to guide the restoration process.

How to Customize the Prompt

To customize, you can emphasize certain areas. If a photo has a severe tear on a face, add a specific instruction like "Pay extra forensic attention to inpainting the face based on logical symmetry." To enforce a specific tone, change the instruction to "Output as a rich, deep sepia monochrome" instead of neutral gray.

For Men / Women / Couples Variations

This prompt is universal and adjusts automatically based on the uploaded subject. Because it specifically forbids altering facial features, bone structure, or expressions, it works equally well for restoration of men, women, couples, and groups without needing to tweak the text.

Common Mistakes

Common issues include AI models "plasticizing" skin, altering facial features to look like different people, or inventively adding modern clothes. This prompt is heavily safeguarded against this with negative constraints, but if it occurs, re-run the generation or emphasize "strict feature preservation" in the input.

Example Results

Expect a clean, sharp, high-resolution version of your old photo with all scratches and stains removed. A faded, washed-out image will gain deep contrast and rich detail, yet it will still look like a vintage photograph rather than a modern digital shot.

Frequently Asked Questions

Can I edit my own photos using ChatGPT or Google Gemini?
Yes! Both ChatGPT and Google Gemini allow you to upload your own photos. You can then provide specific prompts to ask the AI to alter the image, change the background, or apply different artistic styles.
How do I upload an image to ChatGPT or Gemini for editing?
In the chat interface of both platforms, look for the "+" or paperclip/image upload icon next to the text box. Select your photo, type your editing prompt (e.g., "Change the background to a cyberpunk city"), and hit send.
What kind of image edits can these AI models do?
They can analyze your image and generate variations, apply artistic filters, or seamlessly blend your photo with new elements described in your text prompt. You can even ask them to change the lighting or clothing style.
Is there a specific aspect ratio I should use for AI photo editing?
While you can upload an image of any size, it is best to know your target aspect ratio (like 16:9 for desktop or 9:16 for mobile). The AI usually maintains the original aspect ratio of the uploaded photo unless you specifically ask it to expand or crop the image.
Why isn't the AI changing the specific detail I asked for?
AI models can sometimes be unpredictable. If it misses a detail, try to be more specific in your prompt. Use instructions like "preserve the original face exactly as it is but change the jacket to red" to guide the AI more accurately.
❓ Can I use someone else’s photo for AI photo editing?
It’s best to use your own photo for AI photo editing. If you want to edit or transform someone else’s photo, make sure you have their permission or consent first. Respect people’s privacy and use AI responsibly.