Official ComfyUI: Blender blockout to Seedance 2.5
Same idea as LTX Layout-to-Render — 3D controls the camera, AI finishes the shot — but the Seedance 2.5 path, posted by the official ComfyUI account.
On Oct 16 the official @ComfyUI account posted a Blender blockout → Seedance 2.5 workflow, with a graph on comfy.org (https://comfy.org/workflows/3ddb7afa1aa4-3ddb7afa1aa4/). The pattern matches what we covered for LTX: you block the shot in a real 3D tool so camera path and layout are deliberate, then an AI model finishes the look. Here the finisher is Seedance 2.5 instead of the LTX-2.5 Layout-to-Render LoRA.
If you already read our LTX Layout-to-Render Lab post (https://thereelstack.com/lab/ltx-layout-to-render-blender-blocking), treat this as the proprietary-model sibling: same blocking habit, different renderer. Useful when your pipeline already speaks Seedance, or when you want to A/B the same playblast through both paths.
Honesty note: we haven't run the workflow. If the comfy.org page is slow or gated for you, the X post and the workflow URL above are still the canonical pointers. Credit @ComfyUI.
Skills & repos in our directory
- ComfyUI Blender→Seedance 2.5 · ComfyUI, Generation
- seriesmith · ComfyUI, Generation
- FastH3 Omni · ComfyUI, Generation
- ReAngle · ComfyUI, VFX & compositing
- Comfield · Repo, Generation
- FXjutsu · ComfyUI, VFX & compositing
More from the Lab
- Block the shot in Blender, render it in LTX: the Layout-to-Render LoRASpotted, not yet tested
- The top GitHub video repos right now, ranked by stars (Oct 2026)Spotted, not yet tested
- FastH3 Omni goes open: reference-to-video with sound on your own GPU, and the case for letting Opus cut cheap clipsSpotted, not yet tested
- ComfyArt: a local Lovart-style infinite canvas on your ComfyUISpotted, not yet tested
Topics: ComfyUI · Seedance · Open-weight video & audio models