Skip to content

All images and videos deserve a second chance at life

About us

VanceAI was founded in July 2020 by a small team of deep learning researchers in Hong Kong, with offices now in Hong Kong and Japan. We started VanceAI to democratize photo and video restoration through AI.

When we started, pro-grade photo enhancement was locked behind either Photoshop expertise or paid restoration services costing $50 per image. Most people with a damaged family photo had no good option. We thought AI could close that gap, not just for professionals but for anyone with a memory worth preserving.

We scaled from zero to 9.8 million lifetime users, became profitable in 2022, and have built 18+ proprietary AI tools since.

Our tools span photo upscaling, restoration, sharpening, denoising, colorization, and background removal, alongside full-video enhancement. Available through the web app, our desktop application, and our API.

What we believe

Image quality is a craft, not a benchmark. The right model trade-offs come from looking at the output and asking whether the print holds up, whether someone would hang it on a wall, whether the subject still looks like themselves. Academic metrics don’t measure any of those things.
AI should disappear into the workflow. If a tool needs a manual, it isn’t ready. The best work feels like the photo just came out that way.
Restoration, not invention. Our job is to bring back what was in the original, not fabricate what wasn’t there. A restored photo should still be the same photo.
Speed is part of quality, not a trade-off against it. A perfect output that takes 20 minutes per image stops being a tool and starts being a chore nobody uses. We optimize for the lowest reasonable latency, then push the output quality up from there.

Where we’re headed

The next chapter of VanceAI is about making professional restoration available to the people who need it most but can’t access it: families with damaged photos in a drawer, archives and historical societies preserving the past, small businesses without a creative team, photographers handling volume work at margins that don’t allow for hours per image.

We’re investing in better models, faster pipelines, and partnerships with the platforms where this work actually happens: scanning services, photo book makers, memorial software, creative tools.

We want a world where AI disappears into the workflow, where fixing blurry footage is as easy as shooting it sharp to begin with.