Nodes/ComfyTV/Blender Animation
ComfyUI Node

Blender Animation

Turn your Blender scene into a video clip the AI pipeline can chew on

By jtydhr88·Created 3 months ago·Updated about 20 hours ago· 773
Blender Animation
    • video
    force_run_token0
    project_id
    parent_output_id0
    camera
    shadingclay

    "Blender Animation" is the moving-image half of ComfyTV's 3D pair. Where Blender Camera shoots a single still from your scene, this stage renders the whole animation range from a chosen camera and hands you a real video clip - an video output you can drop straight into ComfyTV's video-edit stages, an FX chain, or a compositing pass. If you've built a camera move or animated the scene in the Blender Scene workbench, this is how it leaves the 3D world and joins the AI one.

    It shares a workflow with its sibling: build in Blender Scene, point Blender Camera at a still, point this at the motion. The stage reads the live scene from the blender-web bridge, renders frames from the scene's frame range, and emits the finished clip as a project snapshot. Re-running it re-renders only this stage - the per-node Run model means your video doesn't drag the whole canvas through the queue every time you tweak a camera.

    How it works

    The mechanism is the same remote-render call as Blender Camera, just in "animation" mode. The stage asks the bridge to render the named camera over the scene's frame_startframe_end range, then polls the job, ticking progress up as frames complete ("rendering 37/120" and the like), copies the finished MP4 into ComfyTV's output area, and emits it. Longer scenes mean longer poll times and a deadline that scales with frame count - a 200-frame animation is not a quick run.

    The inputs that matter

    • camera - which camera renders. Picked from the camera picker on the card, not typed.
    • shading - clay (default) is the Workbench driving render - flat form + motion + camera, ideal for AI stages that want clean motion without material noise. full renders with the scene's own engine (Eevee/Cycles, F12-quality).

    The rest - force_run_token, project_id, parent_output_id - is internal plumbing.

    The output

    • video (COMFYTV_VIDEO) - the rendered MP4. Wire it into the video editing / color / keying / FX stages, or use it as reference for a video-generation pass.

    Install

    ComfyUI Manager (search ComfyTV) or:

    cd ComfyUI/custom_nodes
    git clone https://github.com/jtydhr88/ComfyTV
    

    Restart ComfyUI. No extra Python deps, but the blender-web bridge must be running (blender-for-comfytv.bat) or this stage fails before the first frame.

    Where people get burned

    Same bridge dependency as Blender Camera - "Blender bridge is not running" means you forgot to launch the blender-web distribution, and the node card shows the red dot long before you ever hit Run. Beyond that, the trap is expecting a feature film from a 5-frame scene: this renders exactly the scene's frame range, so if you want motion, you animate in the workbench first. And clay is the right default when the output feeds an AI stage - switch to full only when the render is the final look. That one setting is the difference between a motion reference and a finished shot.

    CategoryComfyTV/3D

    Inputs (5)

    NameTypeDefaultDescription
    force_run_tokenINT00–2147483647Internal — bumped on Run to invalidate ComfyUI's input cache.
    project_idSTRINGInternal — populated by the projectStore on the frontend.
    parent_output_idINT00–2147483647Internal — lineage parent set by spawn handlers on the frontend.
    cameraSTRINGScene camera this stage renders. Managed by the camera picker on the card.
    shadingCOMBOclayclay: Workbench driving render (form + motion + camera, for AI stages). full: the scene's own engine, F12.

    Outputs (1)

    NameTypeDescription
    videoCOMFYTV_VIDEO