Do I need the LLM to use this workflow?

No. The LLM (GeminiNode) is optional. You can bypass it and write your own Seedance 2.0–style prompt directly in ByteDance2ReferenceNode. The LLM simply drafts a well-structured prompt from your references and description.

How are the two reference images used?

They ground the content and style the LLM describes and guide Seedance 2.0’s final render. Use one image for the subject and one for style or environment. The ImageConcanate node provides a side-by-side composite to give the LLM clearer context.

Does GeminiNode require an internet connection or API key?

Yes, GeminiNode typically relies on the Google Gemini API. You’ll need a valid API key configured in your ComfyUI environment and a working internet connection. If that isn’t available, bypass the LLM and input prompts manually.

How can I improve consistency and quality?

Use high-resolution, clean references without heavy text/watermarks. Keep the subject well-centered, match lighting between references, and be explicit in your description about camera motion (e.g., “slow pan left”), timing, and mood. Iterate by refining the LLM prompt and re-running.

Workflow

Seedance 2.0 + LLM Prompt Helper

This workflow uses two reference images and a simple text description to automatically generate a detailed Seedance 2.0-style prompt using an LLM. Instead of spending time figuring out complex prompting structure, the workflow analyzes your references and builds the prompt for you, making it much faster and easier to create high-quality results. The included LLM helper is designed to match the style and formatting that works well with Seedance 2.0, helping streamline the entire process while still giving you full control over the final output. Just drop in your references, describe the scene or motion you want, and the workflow handles the prompt generation automatically.

Try on Comfy Cloud Download Workflow

CREATED BY

enigmatic_e 3 months ago

Seedance 2.0 + LLM Prompt Helper streamlines prompt creation for ByteDance/Seedance-style video generation by pairing two reference images with a short scene description. The workflow routes your images through LoadImage and ImageConcanate to form a clear, side-by-side visual context, which the GeminiNode then analyzes along with your text to draft a Seedance 2.0–formatted prompt. PreviewAny lets you inspect and copy the LLM’s output before it’s sent on, while PreviewImage helps you sanity-check your references.

Once you’re happy with the generated prompt, it feeds into ByteDance2ReferenceNode (the Seedance 2.0 runner) to produce the motion result. You can review frames with PreviewAny/PreviewImage (as applicable) and export the final animation via SaveVideo. A MarkdownNote in the canvas documents tips and conventions the LLM uses so the prompt it produces aligns with Seedance 2.0’s expected structure and tags. The result: you spend less time reverse-engineering prompt syntax and more time iterating on visuals and motion.

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.

49a2f3a0811f.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

View all workflows