Why we built B&W colourisation into Cleanr
About three months before we shipped Cleanr, a user emailed me a photograph. It was black and white. A wedding photo from 1987. She'd explained in her message that her grandmother had passed, and she wanted to colour it for the memorial service, but every other app wanted a subscription or charged per use. That email sat in my inbox for a week.
The moment we realised we were missing something obvious
That email wasn't unique. After we launched Cleanr with object removal, background swapping, and sky replacement, we started getting messages from users asking if we could add one thing: the ability to bring colour back to old photographs. Not glossy filters. Not fake saturation overlays. Real colourisation.
We're a small team at MRVL. We had 22 tools already in the app. Adding another feature meant choosing what mattered most. But here's what struck me: people weren't asking for another trendy preset or a subscription tier. They were asking for something practical. Something that solved an actual problem in their lives.
Old photos matter. Family photos, historical images, archives, memories. And the people who care about them aren't professional photographers. They're everyday folks who want to restore something meaningful without complexity or cost.
The technical choice that shaped how we built it
We could have gone two directions. One was to layer a fancy neural network on top. The other was to use a colour lookup table (LUT) built on real colour data. We chose the second route. Specifically, we implemented CIColorCube, which sits inside CoreImage on iOS and Android.
Why? Because it works. CIColorCube uses a mathematically sound approach to map greyscale values to naturalistic colour. It's not trying to guess what your grandmother's dress was. It's producing colours that look real because they're based on colour distributions that actually exist in photographs.
The upshot is this: when someone runs colourisation on a black and white photo in Cleanr, they get a result in seconds that looks believable. Not perfect. But honest. And crucially, it doesn't require an internet connection, a wait time, or a credit balance.
It had to fit into a bigger vision
Here's something we've always believed about Cleanr: it shouldn't be a fragmented experience. You shouldn't need five different apps to clean, restore, and enhance a photo. That's madness. So when we added colourisation, we didn't just drop it in as a standalone button. We threaded it into the restoration pipeline.
You take a black and white photo, you remove dust and artefacts with our JPEG cleanup tool, you bring back colour with colourisation, then you adjust warmth and saturation to taste. Or you batch process ten photos at once on the Plus tier. It flows.
That's a departure from how a lot of photo editors work. They're built around individual effects. We're trying to build Cleanr around what people actually do when they're trying to bring an old photo back to life.
Why it matters that we didn't charge extra for it
Free users get two colourisations a day. Plus subscribers get unlimited. That's it. No pop-up asking you to watch an ad. No "credit" system with opaque math. No dark patterns.
I've watched a lot of photo editor companies build themselves around freemium models that feel almost hostile to the free user. They're not making tools. They're optimising for conversion friction. I didn't want Cleanr to be that.
When the user with the 1987 wedding photo opens Cleanr, she can colourize that image right now, today, for free. She might hit her daily limit after two photos. But she won't feel trapped. And if she wants unlimited, the pricing is transparent and fair.
What we learned along the way
The most interesting discovery wasn't technical. It was about who uses colourisation. We expected historians and archive workers. We got a lot of those. But we also got faith creators preserving family histories, small business owners restoring product photos for e-commerce, and people working on genealogy projects.
One user submitted a batch of fifty photos from an estate sale, all monochrome, all from the 1950s. She was documenting them before selling them off. Another user colourised photos of her grandparents to include in a tribute video for their anniversary.
That broadened how we think about what restoration means. It's not just technical cleanup. It's the work of memory. And every tool in Cleanr needs to respect that.
When you're building a photo editor, it's easy to get caught up in what's trendy or what competitors are launching. Colourisation isn't trendy. It's needed. The question we asked ourselves was simple: if someone brings us a black and white photograph that matters to them, can we help them without making them jump through hoops? That's what we built.