Old family photos carry a lot of history, but the physical prints they live on don't age well — colors fade, scans come out soft, and black-and-white shots can feel distant compared to how people actually remember the moment. AI restoration tools can't recover information that was never captured, but they can do a surprising amount with what's already there.
Step 1: colorize
Start with Colorize. The model looks at the content of a black-and-white photo — skin tones, sky, foliage, fabric — and infers plausible, natural-looking color. It won't know the exact color of your grandfather's shirt, but it produces a believable, warm result that usually feels much more alive than the grayscale original.
Step 2: sharpen
Scanned prints are often a little soft, especially if the original print itself had degraded. Run the colorized result through Unblur to recover edge detail that a flatbed scanner or phone camera softened.
Step 3: upscale (optional)
If you want to print the photo larger than the scan's native resolution supports, finish with Upscale at 2x or 4x. Doing this last (after colorizing and sharpening) means the upscaler has more real detail to work from.
A note on accuracy
Treat AI colorization as an artistic restoration, not a historical record — the model is filling in plausible color, not recovering a color that was actually captured. For photos where color accuracy matters (rare, but it happens), pair the AI pass with any color notes or memories you have and touch up manually in the editor.
Putting it together
Colorize → Unblur → Upscale is a good default order for most old scans. Each step is a separate tool, so you can stop after any one of them if that's all a particular photo needs.