Hacker Newsnew | past | comments | ask | show | jobs | submit | poznerd's commentslogin

I’ve been working on a side project called Forevi for the past ~2 months.

It started as a personal thing when I tried to restore some old family photos, especially for my grandpa. He didn’t want a material gift for his birthday - he’s one of those people who says they don’t need anything, but actually cares a lot about memories. Most tools I tried either changed faces or added weird filters, making everything look overly “AI-smooth,” which kind of defeats the point when you’re dealing with real moments.

So I built Forevi with a simple focus: restoring old photos without changing faces or geometry, adding natural color instead of filters, upscaling images so they’re actually printable, and optionally animating photos with close-to-real motion and sound. It’s meant mostly for real family archives and any old and/or damaged photos.

I don’t expect to make a lot of money from this. There are paid plans mainly to cover generation costs (no crazy 300% ROI like some similar services). This is very much a side project.

I’d really appreciate honest feedback on what feels right, what feels off, and what you’d improve or remove.

Thanks, Daniel


Nice. Some feddback:

It's difficult to parse the title in HN. You "use AI fix the photos" instead of "use AI to ruin the photos". I expected an expert system like it was 1999.

I'd still try to add some ROI to cover unexpected unexpected, like card frauds.

I expected to get the full image when I click the samples in the middle. It only increase the size a little.

The example with the two tranways and a tower (London?) looks horrible in my browser. It's very nice when I open it alone in a new tab, but the small video get's a lot of artifacts in my browser.


Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: