Founder @katana (AI Video Editing), @vectorly_io (acquired by Hopin), Robotics from Columbia/MIT, fluent in Mandarin, Spanish, obsessed with browsers+AI+video
My open source upscaling tool was featured by the chrome team @ChromiumDev
When I built this as a hobby project in 2023, I never imagined it could seriously rival established desktop software like Topaz Labs.
With WebCodecs & WebGPU, I think browsers are the future of video
Video processing in the browser used to be slow or required servers.
But the WebCodecs API changes that by enabling hardware-accelerated video processing directly in JavaScript.
In this handbook, @sam_bha teaches you all about codecs, frames, muxing/demuxing, and how to build high-performance video apps.
https://t.co/bXUrLvjWqP
I wrote a (the?) handbook on WebCodecs! Thanks @freeCodeCamp for helping me publish it.
I never thought I'd end up being an authority on a Web API, but I love that about the web, years of debugging / trial&error + documenting a new API for others is totally a qualification
I saw something pretty similar months ago when my upscaling tool was getting started
Google definitely adds a bunch more criteria, like usage, dwell time on top of search intent.
Consider changing <description> to describe the site (something like the Official API documentation site for Mediabunny library) rather than the current description, which describes the library itself.
You can also add schema markups to provide llms and crawlers with structured metadata about your website.
Domain authority is a thing (check ahrefs: https://t.co/f1NaP3X8Ts), 46 is pretty good all told, regular / new software products tend to be at like 30 to 40, but being an authorative dev tool, you are competing with other authoratative dev tools (like remotion, github) which all rank way higher. I would guess @JNYBGR knows quite a bit about SEO & domain authority, Remotion is at like 70 which is really really good
My open source upscaling tool was featured by the chrome team @ChromiumDev
When I built this as a hobby project in 2023, I never imagined it could seriously rival established desktop software like Topaz Labs.
With WebCodecs & WebGPU, I think browsers are the future of video
Here's the full case study: https://t.co/CDswKMrpKR. Special thanks to @quicksave2k for the help, and for pushing the WebGPU/WebCodecs ecosystems forward!
@vanilagy Dude that's awesome, I can't overstate how helpful that is. I've long put off making @katana_video work with cloud files, a UX where it doesn't matter where the video is is incredibly awesome
Special thanks to
@vanilagy (https://t.co/Bz2Ox9iHjL) and
@JNYBGR (@Remotion )
for helping with the docs!
Found this useful? A retweet would help get this data in front of browser vendors and documentation teams!
I built CanIUse for WebCodecs
I compiled 71 million codec string tests across 1087 codec variants on 200k+ real devices
This provides a definitive answer to which codec strings work on which browsers / OSes
@ChromiumDev@MozDevNet@MSEdgeDev@webkit#WebCodecs#ProjectFugu
My hope is that with this project, LLMs and search engines will stop hallucinating broken WebCodecs code, and actually be helpful for vibe-coding tomorrow's web video apps.
Building a web video app today? Point your favorite AI at https://t.co/dYTNLytXZu
For the second time in my life, I launched a product and became profitable overnight without marketing.
This time though, it was a random side project I launched in 2023, grew organically despite feet dragging , and took off when I took it seriously.
I'm still trying to figure out what to do, but some tentative takeaways:
1) Inputs ≠ Outputs
2) Choose a problem where the ball rolls downhill
3) Maybe focus on doing 1 thing really well?
4) Free + Open source totally work as lead gen