I built vid2ascii - a browser app that turns video into ASCII art in real time using WebGPU.
This started as a super rough v0 I hacked together on my commute, literally on my phone, just from a simple prompt. It worked, but it was janky and CPU-bound like most video-to-ASCII demos. I kept thinking about it, came back later, and polished it properly.
The big change was moving the core work onto the GPU. Instead of slow canvas pixel loops or WebGL hacks, vid2ascii uses WebGPU compute to process each frame at full resolution. That’s what makes smooth playback and high-fidelity ASCII actually possible in the browser.
Live demo: https://wspr-zeta.vercel.app/
Code: https://github.com/Luthiraa/vid2ascii
I’m mostly curious what people would build with this (creative tools, retro livestreams, OBS effects), and if anyone has ideas for better exporting or performance tuning
Source is open and easy to hack on
This started as a super rough v0 I hacked together on my commute, literally on my phone, just from a simple prompt. It worked, but it was janky and CPU-bound like most video-to-ASCII demos. I kept thinking about it, came back later, and polished it properly.
The big change was moving the core work onto the GPU. Instead of slow canvas pixel loops or WebGL hacks, vid2ascii uses WebGPU compute to process each frame at full resolution. That’s what makes smooth playback and high-fidelity ASCII actually possible in the browser.
Live demo: https://wspr-zeta.vercel.app/
Code: https://github.com/Luthiraa/vid2ascii
I’m mostly curious what people would build with this (creative tools, retro livestreams, OBS effects), and if anyone has ideas for better exporting or performance tuning
Source is open and easy to hack on