CCPP Vids
Videos written as code, rendered on the CPU, with no generative model in the picture.
Planned: .vid.lua projects that describe scenes, timing and speech, rendered deterministically on the CPU. No AI anywhere except the voice.
- rendering, no GPU required
- CPU
- scene files, versioned in git
- Lua
- generative video models
- 0
- voice model, for speech only
- 1
Product demos, tutorials and changelogs age with every release. CCPP Vids makes a video a source file: Claude edits the scene, the renderer produces the same frames every time, and the video is rebuilt with the product.
What we are building
Videos as code
Scenes, layers, timing and captions in a .vid.lua project that diffs, reviews and versions like the rest of the repository.
Rendered on the CPU
No GPU and no cloud render farm. A VPS renders the same frames a laptop does.
AI for the voice only
Speech is the one part a model generates. Every frame is computed.
Status
CCPP Vids is in its first phase: measuring what a CPU renderer can do on the hardware our users actually have. Create an account to hear when it opens.
Be the first to know
Join the waitlist with a free account.
Works well with CCPP Vids
Every productCCPP MC
Claude builds Minecraft: servers, mods, maps and packs, Java and Bedrock, and sees what it made.
CCPP Lua
The harness Claude Code was missing: one Lua chunk does a whole step of work.
ccpp shell
One Bash, the same on every operating system, with its core commands written in Go.
One account. Every tool Claude needs.
Free products come with a free account. Builder unlocks the rest, the harness first.