ComfyUI Wan dance test renders a 30-second clip in 70 minutes
tostane · reddit · 2026-07-22
ComfyUI Wan dance test took 70 minutes for a 30-second clip
The poster tried a Wan dancer workflow in ComfyUI for a K-pop fast-dance clip.
- Using an image of 3 girls did not work well; the model seemed to prefer 1 girl.
- A second test with 1 girl at medium speed worked better.
- Rendering a 30-second dance video took 1 hour 10 minutes.
- If you try it, the poster recommends editing the workflow and adding a Trim Audio Duration node.
More from coding & agent
- Agents need realistic environments for API keys, security hurdles and RL hacks — 1a3orn · 2026-07-22
- DataFlow-Harness uses MCP and DAG edits to build editable LLM pipelines — PekingUniversity · 2026-07-22
- AI video’s bottleneck is workflow, not generation quality — aftahi_ai · 2026-07-22
- Andrew Ng’s graph-based agent design claims a 90% efficiency gain — Roger_M_Taylor · 2026-07-22
- Human approval gates are becoming standard for production AI agent workflows — Roger_M_Taylor · 2026-07-22
- Hermes Agent users can cut tokens 30–50% by routing background tasks to cheaper models — Roger_M_Taylor · 2026-07-22