I got tired of DaVinci Resolve crashing my laptop, so I built a color grader that runs in a browser tab

A few months back I was trying to color grade some footage for a small project and DaVinci Resolve kept choking on my machine. Nothing fancy, just a mid-range laptop, but Resolve wants a real GPU and mine is... not that. Premiere wasn't much better. I'd sit there watching the preview lag two seconds behind every slider move, and at some point I just closed both apps and started sketching out something simpler.
That "something simpler" turned into Featherline.
the actual problem I was solving
It's not that DaVinci or Premiere are bad tools. They're genuinely great if your hardware can keep up. The problem is a lot of people's hardware can't — students on old laptops, people editing on a phone because that's the only screen they have, hobbyists who don't want to spend on a new rig just to color grade a few clips. That's a much bigger group than I think the "pro tools" market accounts for.
So the goal wasn't "build a worse Resolve." It was: build something that does the core color work — wheels, curves, scopes, that kind of thing — but doesn't assume you're running a workstation.
what actually turned out to be hard
Getting real-time grading to feel responsive without a dedicated GPU was the annoying part. Every adjustment (lift/gamma/gain, curve points, saturation) needs to redraw the preview fast enough that it doesn't feel like you're dragging a slider through mud. On a phone or an old laptop that's a real constraint, not a theoretical one — you notice the lag immediately.
That's basically why the render engine ended up being written in Rust and compiled to WASM instead of just doing the math in JS. Running the actual color transforms outside the JS main thread made a bigger difference to low-end performance than any other single decision I made. Export runs through a separate API layer rather than trying to do everything client-side, which also kept the browser from choking on bigger files.
I ended up spending way more time than expected just on making that rendering pipeline lightweight rather than adding more features. Which is a weird thing to be proud of, but cutting things down so it runs smoothly on low-end hardware ended up being the actual hard engineering problem, not the color science part.
The other thing I didn't expect to matter as much: making it work reasonably well on mobile browsers specifically, not just "responsive design" mobile, but actually usable with touch input for things like curve adjustments where precision matters.
where it's at now
Featherline is live at featherline.pages.dev. It's free, no watermark on exports, and runs entirely in the browser — no installs, no account wall to try it. You can grade photos or video clips with curves, color wheels, scopes, and a handful of presets/LUTs.
I haven't added any paid tier yet on purpose. Right now I care more about it actually being useful to people who couldn't otherwise afford or run this kind of tool, and about finding what's broken, before I think about monetization at all.
If you've got an old laptop or you're stuck editing on a phone and wanted to try color grading without installing a 15GB app, I'd genuinely like to know if it holds up on your device. That kind of feedback is way more useful to me right now than anything else.
Link: https://featherline.pages.dev
Happy to answer questions about the Rust/WASM setup or any of the tradeoffs if anyone's curious — always interested to compare notes with people who've built similar performance-heavy tools for the browser.