My Co-Founder works for Mr.Beast, I edit in Hollywood for a living...(WebGPU talk in the post)
Quick intro. I edit in Hollywood for a living. My co-founder works for Mr. Beast. Between us we have spent a lot of years staring at timelines, and we both believe the same thing: video is becoming the main way people talk to the world, and the tools to make it should be free and they should be fast.
So we decided to build a free, open source video editor on WebGPU.
We have watched projects like this die before. We think there are three reasons why that is, and we want to build around them.
- The wrong people build them. Most of these start with web developers, and no shade at all, but we have seen too many struggle with the timeline. Game developers tend to grasp this space faster. Frame timing, a render loop, GPU state, scheduling. It is closer to a game engine than a CRUD app.
- Audio is the secret killer. Everyone obsesses over video and compositing, then audio quietly takes the whole thing down. In a real timeline editor audio has to be the master clock. Sync video to audio, not the other way around, or it never feels right. Audio is also notoriously hard in the browser and theres more limitations to count.
- The UI. Editors share muscle memory built on Adobe, CapCut, Resolve, FCP. Deviate too far from that and you get zero adoption no matter how good the engine is.
We already have a working prototype built on Mediabunny and PixiJS (not vibe coded). It does the job, but it was never meant to be the real thing. What we actually want to build is a proper compositing engine designed from the ground up for video playback in the browser, not a render library we bent into a timeline. WebGPU finally makes that feel realistic. Rive's new GPU layer for 3D, shaders, and custom effects just hit early access and is a good signal that the pieces are there now.
Here's an open question we keep going back and forth on: 2D or 3D. We are genuinely undecided. The instinct is to stay 2D since that is what most editing is. But at MrBeast there is a ton of round-tripping into After Effects, and AE is really 2.5D, 2D objects living in a 3D space. So we are leaning toward maybe building a 3D compositor that is optimized for 2D rendering by default, then lets you switch on real 3D capabilities when a shot needs it. Curious what people here think about that tradeoff, because it shapes basically every decision underneath.
We would probably ship the editor itself as an Electron app. As an editor I want to minimize distractions, and a focused desktop window beats a browser tab buried in everything else. But the engine underneath stays web, stays open. We also just want to focus on Chrome for a lot of reason (I know the pros and the cons).
So that is the plan. Tear the experimental parts out of the prototype, keep the bones, build the real thing free and open source.
If this sounds like your kind of problem, we are putting together a list of people who might want to work on it. Shoot me a DM. Or just roast this idea in the comments.