Every family has a shoebox of old photographs -- faded portraits from decades past, cracked prints from a grandparent's wedding, water-stained snapshots of childhood vacations. These images are irreplaceable memories, yet time, humidity, light exposure, and simple handling wear them down. Scratches appear across faces. Colors shift to yellow and sepia. Entire sections crumble or flake away.
For generations, restoring these photos required a skilled retoucher spending hours in Photoshop, carefully cloning pixels, repainting missing details, and correcting color shifts by hand. The results could be impressive, but the process was expensive, slow, and inaccessible to most people.
AI photo restoration has changed everything. Modern neural networks trained on millions of images can now analyze damage patterns, reconstruct missing detail, remove scratches, correct color casts, and sharpen blurry areas -- all in seconds rather than hours. What once cost hundreds of dollars at a professional retouching studio can now be done on your own computer with a few clicks.
In this guide, you will learn exactly how to restore old photos with AI using Rangy, a desktop app for Mac and Windows that brings professional-grade AI restoration directly to your machine.
What AI Photo Restoration Can Fix
Before diving into the steps, it helps to understand what kinds of damage AI restoration handles well. Modern restoration models have been trained on vast datasets of damaged and pristine photo pairs, so they understand the difference between intentional image content and various forms of degradation.
AI excels at removing scratches, correcting fading, and reducing noise because these are consistent, learnable patterns. Reconstructing large missing areas is more challenging -- the AI will generate plausible content to fill gaps, but for critical details like faces, you may want to combine AI restoration with manual touch-ups for the most faithful result.
Step-by-Step: Restore Old Photos with Rangy
Rangy includes a dedicated Restore fast action in its Pro tab, purpose-built for repairing old and damaged photographs. Here is how to use it from start to finish.
Download and install Rangy
Head to rangy.ai/download and grab the installer for your platform (Mac or Windows). Installation takes under a minute. Once launched, enter your Replicate API key or Kie.ai API key in the Settings tab to connect Rangy to the AI processing backend.
Scan and open your old photo
If you are working with a physical print, scan it at the highest resolution your scanner allows -- 600 DPI or higher is ideal. Save the scan as a PNG or TIFF to preserve quality. Then open the image in Rangy by dragging it onto the app window or using File > Open.
Select the area to restore
Use the selection tool to highlight the area you want to repair. For most old photos, select the entire image. If damage is concentrated in a specific region, you can select just that portion for a more targeted restoration.
Use the Restore fast action
Switch to the Pro tab in Rangy and click the Restore fast action button. This sends your selection to an AI restoration model that has been specifically trained to identify and repair photo damage. The model analyzes the entire image, distinguishes genuine content from degradation artifacts, and produces a clean, repaired version. Processing typically takes 10-30 seconds depending on image size.
Upscale the restored result (optional)
Old photos are often small and low-resolution, especially after scanning. Once the damage is repaired, use the Crystal upscaler to increase the resolution by up to 10x. This adds genuine detail and sharpness rather than simple interpolation, making the photo suitable for large prints or modern high-resolution displays. Rangy also offers the Magnific upscaler for an alternative approach with different creative characteristics.
Save your restored photo
Export the final image in your preferred format -- PNG for maximum quality, JPEG for smaller file sizes, or WebP for web use. Your original file remains untouched, so you can always return and try different restoration approaches.
For severely damaged photos, try running the Restore action twice. The first pass handles the major damage, and a second pass can clean up any remaining artifacts for an even smoother result.
Tips for Best Results
AI restoration works best when you give it the most information to work with. Here are practical tips to get the highest quality output from your old photo restorations.
Scan at the highest DPI possible
A 600 DPI scan captures far more detail than a 300 DPI scan, and that extra data gives the AI more to work with. If your scanner supports 1200 DPI, even better. More pixels mean more context for the restoration model to understand what the original image looked like before damage occurred.
Restore before colorizing
If you plan to both restore and colorize a black and white photo, always restore first. Removing scratches, cracks, and stains from a grayscale image is simpler for the AI because it does not need to reason about color. Once the damage is repaired on the clean grayscale version, colorization will produce much more consistent and natural results.
Use Crystal upscaler after restoration
The Crystal upscaler works best on clean images. Running it on a damaged photo will upscale the damage artifacts along with the content. By restoring first and upscaling second, you get a sharp, high-resolution image free of scratches and stains. The Crystal upscaler supports scale factors up to 10x, which can turn a small 500x400 scan into a 5000x4000 pixel print-ready image.
Batch process family albums
If you have a stack of old photos to restore, Rangy supports batch processing. Scan all your photos, open them one at a time, apply the Restore fast action, and save. The process takes only seconds per image once you have the workflow down, so you can work through an entire family album in an afternoon.
Keep your originals
Always keep the original scans untouched. AI restoration is non-destructive in Rangy -- the result is placed as a new layer -- but it is good practice to archive your raw scans separately. As AI models improve over time, you may want to re-process your originals for even better results in the future.
How Much Does AI Photo Restoration Cost?
Rangy uses a pay-per-use model for AI processing. You are not locked into a monthly subscription for usage credits -- you only pay for what you generate. The AI processing runs through Replicate or Kie.ai, and you connect your own API key.
Here is what typical operations cost:
- Photo restoration: Varies by model and image size, generally a few cents per image
- AI image generation (e.g., Seedream 4.5): Approximately $0.04 per image
- Upscaling with Crystal: Varies by resolution and scale factor
For a typical restoration workflow -- restore plus upscale -- you are looking at well under a dollar per photo. Compare that to professional retouching services that charge $25-$100 or more per image, and the value is clear. Even restoring an entire family album of 50 photos would cost only a few dollars in API usage.
You can check current model pricing on the Rangy pricing page and in the documentation for a detailed breakdown of costs per model.
Rangy vs. Online Photo Restoration Tools
There are several online photo restoration services available, but a desktop application like Rangy offers distinct advantages for anyone serious about restoring their photo collection.
| Feature | Rangy (Desktop) | Online Tools |
|---|---|---|
| Privacy | Images stay on your machine | Uploaded to third-party servers |
| Upload limits | No file size restrictions | Often 5-25 MB max |
| Resolution | Full resolution, up to 10x upscale | Often downscaled output |
| Batch processing | Process multiple photos efficiently | One at a time, re-upload each |
| Multiple AI models | Choose from several models | Typically one fixed model |
| Affinity/Photoshop integration | Works as a plugin | Not available |
| Cost transparency | Pay-per-use, own API key | Monthly subscriptions |
The privacy advantage alone is significant. Old family photos are deeply personal, and uploading them to unknown servers raises legitimate concerns. With Rangy, the image data is sent directly from your machine to the AI provider (Replicate or Kie.ai) through your own authenticated API connection, and nowhere else. There is no middleman storing your family photos.
For photographers, archivists, or anyone working with high-resolution scans, the absence of upload limits and resolution restrictions is another critical difference. Online tools frequently compress your uploads and cap output resolution. Rangy processes at full resolution and lets you upscale beyond the original with the Crystal or Magnific upscalers.
Restore Your Old Photos Today
Download Rangy for Mac or Windows and bring your family memories back to life with AI-powered photo restoration.
Download RangyConclusion
AI photo restoration has made it possible for anyone to repair, enhance, and preserve old photographs without professional retouching skills or expensive software subscriptions. Scratches, cracks, fading, water damage, and yellowing that would have taken hours of manual work can now be corrected in seconds.
With Rangy, the entire workflow -- from opening a scanned photo to restoring damage and upscaling the result -- happens in a single desktop application. After restoring, you can upscale the image to increase its resolution without losing the restored details. Your images stay private, you pay only for what you use, and you get access to multiple AI models so you can choose the best approach for each photo.
Whether you are restoring a single treasured portrait or digitizing an entire family archive, AI-powered restoration through Rangy gives you professional results at a fraction of the traditional cost. Download Rangy and start bringing your old photos back to life.