Seedance2.0 Real Human: Image to Video is a compliant, identity-verified image-to-video workflow built with ByteDance partner nodes in ComfyUI. It guides you through creating verified real-human assets first, then generating a personalized video that transfers motion from a reference video to your identity photo. The verification step is handled by ByteDanceCreateImageAsset and ByteDanceCreateVideoAsset, which turn your uploaded portrait and reference video into reusable asset IDs and a shared Group ID that ties same-person assets together for backend recognition.
Once assets are approved and IDs are available, ByteDance2ReferenceNode performs the image-to-video generation using those IDs (instead of raw media) to meet the provider’s legal requirements. PreviewAny lets you quickly check intermediate info or the final clip, and SaveVideo writes the result to disk. LoadImage and LoadVideo handle local file inputs, while the MarkdownNote provides in-graph guidance. The workflow also includes a Save IDs for future reuse section so you can store and reuse your Group ID and asset IDs, avoiding repeat uploads for the same person. Asset limits apply (images up to 9, videos up to 3, audio up to 3).
API
從程式碼使用此工作流
每個 Comfy 工作流都是一個 JSON 圖。下方的 payload 就是此工作流，與 ComfyUI 執行時完全相同 — 你可以從此 URL 取得、放入版本控制，或在 ComfyUI 載入並逐節點執行。
使用 Comfy SDK 以 TypeScript 或 Python 執行。相同程式碼可用於 Comfy Cloud 或你自架的 ComfyUI — 只需更改 base URL。
// 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");SDK 需使用 API 格式的工作流：請在 ComfyUI 開啟此工作流，並使用 檔案 → 匯出工作流（API）。
常見問題
