Hey HN,
For nearly two years, I’ve been building an AI-native solution for programmatic video editing after running into the many limitations of FFmpeg when automating content creation.
The result is @diffusionstudio/core – a browser-based video editing engine powered by WebCodecs. The library is almost entirely written from scratch, with just a single dependency (49KB gzip).
Since rendering and encoding happen mostly on the GPU, client-side, it’s blazingly fast. And with no need for backend rendering infrastructure, integration into any website is simple and lightweight.
One ed-tech startup used it to turn study materials into engaging videos and scaled to $200K MRR in just 9 months using the engine.
The result is @diffusionstudio/core – a browser-based video editing engine powered by WebCodecs. The library is almost entirely written from scratch, with just a single dependency (49KB gzip).
Since rendering and encoding happen mostly on the GPU, client-side, it’s blazingly fast. And with no need for backend rendering infrastructure, integration into any website is simple and lightweight.
One ed-tech startup used it to turn study materials into engaging videos and scaled to $200K MRR in just 9 months using the engine.
Some of its key features are:
Would love for you to check it out and hack away with it!
Try it for free https://github.com/diffusionstudio/core
A star would mean the world!