Do I need OpenPose or a skeleton extraction step?

No. Wan-Animate-2 performs direct motion transfer from the driving video frames. The workflow does not require keypoints or pose preprocessing.

How do I keep the character identity consistent?

Use a clean, well-lit reference image with minimal occlusion and a simple background. Keep the aspect ratio consistent, and avoid prompts that describe a different character or facial features that conflict with your reference.

How do I match the timing and duration of the driving video?

GetVideoComponents provides fps for the driving clip—pass that to CreateVideo so playback matches the source timing. To retime, adjust the fps in CreateVideo (higher fps = faster playback, lower fps = slower).

My clip is long. How do I process the whole video?

Native loop support is still under review. For now, duplicate the "Motion Transfer (Wan Animate 2)" subgraph: enable the bypassed copy (Ctrl+B), ensure its prompt and size match the first, connect its image output to BatchImagesNode, then copy-paste (Ctrl+C, Ctrl+Shift+V) enough segments to cover the full driving video.

Wan Animate 2: Motion Transfer

This ComfyUI workflow demonstrates end-to-end motion transfer using Wan-Animate-2 (by Wan-Video/Alibaba): a single reference image (your character) is animated to follow the motion in a driving video, without any pose keypoint or skeleton extraction. The core of the graph is the "Motion Transfer (Wan Animate 2)" subgraph powered by custom Wan-Animate-2 nodes (UUIDs: 11706f8a-428d-4ef9-b24f-863f651c1b0b, 79280513-53b2-4417-93f8-f3e116bb6736, b798a87e-23ec-483d-ab9b-6af5b298d0ad). These nodes accept your reference image, the per-frame motion signal from the driving video, and a text prompt that guides background look and camera feel. Because motion is learned directly from frames, setup is fast and robust—no OpenPose or preprocessing passes.

LoadImage provides the character identity, while LoadVideo and GetVideoComponents extract the driving frames and fps. BatchImagesNode replicates the reference image so each video frame gets its own paired identity input. GetImageSize and ComfyMathExpression are used to keep resolution and aspect ratio consistent between inputs and outputs. PreviewAny lets you inspect intermediate frames, and CreateVideo + SaveVideo assemble the generated frames back into a video using the original fps. For longer clips, you manually duplicate the Motion Transfer subgraph (native looping support is pending upstream) and connect it in sequence so the pipeline can process the entire driving video.

API

Use this workflow from code

Every Comfy workflow is a JSON graph. The payload below is this workflow, exactly as ComfyUI runs it — fetch it from the URL, keep it in version control, or load it in ComfyUI and run it node by node.

9394f9968da3.json

Fetching workflow JSON…

Run it from TypeScript or Python with the Comfy SDK. The same code targets Comfy Cloud or a ComfyUI you host yourself — only the base URL changes.

// Install (beta)
npm i @comfyorg/sdk

// Run this workflow (TypeScript)
import { Comfy } from "@comfyorg/sdk";

const client = new Comfy({ apiKey: "comfyui-..." });
const wf = await client.workflows.fromFile("workflow_api.json");
const job = await client.run(wf);
await job.getOutputs("<output-node-id>")[0].toFile("output.png");

The SDK takes a workflow in API format: open this workflow in ComfyUI and use File → Export Workflow (API).

Comfy Cloud API access requires a plan with an API key.

SDK docs Get an API key

FAQ

Frequently Asked Questions