streaming_ code_ world_ model_ generate
1 MCP server offers this tool
What it usually takes
chunks– in 1 of 1scene– in 1 of 1seed– in 1 of 1
No server marks it as only reading: treat it as able to change data.
From each publisher's own tool list. We never call a tool.
Servers
| Server | Inputs | Read-only? |
|---|---|---|
Streaming Code World Model co.huggingface.hugging-apps/streaming-code-world-model Stream a world-model video chunk by chunk from a released Code World Model scene. Each 17-frame chunk (0.71 s at 24 fps) is generated causally from the start image, its caption, the Depth + Semantic-ID proxy up to that chunk and the previous chunks (KV cache, 6-chunk attention span), with the 4-step DMD streaming model. Returns: (latest-chunk preview frame, final mp4 path) -- yielded progressively. | chunks,scene,seed | Not stated |