Can I run Seedance 2 in ComfyUI?

Yes. Community-developed custom nodes connect ComfyUI to the Seedance API for both text-to-video and image-to-video generation. Install via ComfyUI Manager or manual git clone, add your API key, and you can generate Seedance videos directly inside node-based workflows.

What GPU do I need for Seedance ComfyUI workflows?

For API-based workflows, any GPU that runs ComfyUI is sufficient because inference happens on remote servers. If you add local post-processing nodes like upscaling or frame interpolation, 8GB+ VRAM is recommended. For potential future local inference, expect 24GB+ VRAM requirements.

Is Seedance ComfyUI free to use?

ComfyUI and the custom nodes are free and open source. You need a Seedance API key from BytePlus, fal.ai, or Replicate. BytePlus offers 2 million free tokens for new accounts. After that, costs are per-generation based on resolution and duration.

How does ComfyUI compare to Dreamina for Seedance?

Dreamina is easier for beginners with a simple web interface. ComfyUI offers workflow chaining, batch processing, integration with ControlNet and IP-Adapter, local post-processing, and full automation. ComfyUI is better for power users and production pipelines.

Can I use custom nodes with Seedance in ComfyUI?

Yes, that is the primary advantage. You can chain Seedance generation with any ComfyUI node: upscalers, frame interpolation, ControlNet, IP-Adapter, face restoration, style transfer, and thousands of community nodes.

Does Seedance ComfyUI support audio generation?

Seedance 2.0 supports audio in its native API. In ComfyUI, you can enable audio output in the Seedance node settings if your API provider supports it. Alternatively, chain a separate audio generation node like Stable Audio or MusicGen after your video output.

What is the maximum resolution for Seedance ComfyUI?

The maximum resolution depends on your API provider. Most support up to 1920x1080 (1080p) for Seedance 2.0. You can generate at lower resolution and use local upscaling nodes in ComfyUI to reach 2K or 4K output.

Can I batch process videos in Seedance ComfyUI?

Yes. ComfyUI has a built-in queue system. Load multiple prompts from a text file or CSV, connect them to the Seedance node, and queue all generations. Each prompt runs sequentially with automatic saving. This is one of ComfyUI's biggest advantages over Dreamina.

Where do I get Seedance model files for ComfyUI?

Current Seedance ComfyUI workflows use API-based inference, so no local model files are needed. If ByteDance releases open weights in the future, they will be available on HuggingFace. For now, you only need an API key from BytePlus, fal.ai, or Replicate.

Is ComfyUI better than the Seedance API directly?

ComfyUI wraps the same API but adds a visual workflow builder, node chaining, batch processing, and integration with local processing tools. Direct API is better for developers building custom applications. ComfyUI is better for artists and creators who want visual control without writing code.

COMFYUI WORKFLOW NODES

Seedance 2 ComfyUI Guide

Build node-based AI video generation workflows with Seedance 2 inside ComfyUI. From installation to advanced ControlNet pipelines, batch processing, and GPU optimization — the complete technical guide for power users who want full control over every parameter.

Start Installation → Jump to Workflows

INTRODUCTION

What Is ComfyUI?

ComfyUI is an open-source, node-based graphical interface for running AI models. Instead of typing commands or clicking buttons in a web app, you connect visual nodes in a graph — each node performs one operation, and data flows between them through connections.

Node-Based Architecture

Every operation is a discrete node: loading a model, encoding a prompt, running inference, saving output. You drag connections between nodes to build a pipeline. This makes complex workflows visual and reproducible. You can save an entire workflow as a JSON file and share it with anyone.

Originally built for Stable Diffusion image generation, ComfyUI has expanded to support video models, audio generation, 3D rendering, and more. The community has created over 5,000 custom node packs covering everything from face swapping to real-time previews.

Why Creators Choose ComfyUI

Reproducibility: Save and reload exact workflow configurations
Composability: Chain multiple models in a single pipeline
Automation: Queue hundreds of generations with different parameters
Efficiency: Smart memory management keeps only active models in VRAM
Community: Thousands of custom nodes extend functionality
Free & Open Source: No subscription, no cloud dependency

WHY COMFYUI

Why Use Seedance with ComfyUI

Dreamina is the easiest way to generate Seedance videos. But ComfyUI unlocks capabilities that no web interface can match. Here are the three biggest reasons to run Seedance inside ComfyUI.

⚙

Custom Workflows

Build multi-step pipelines that go far beyond generate-and-download. Chain Seedance with upscalers, frame interpolation, color grading, watermark removal, and format conversion — all in a single automated flow. Generate a video, upscale every frame with Real-ESRGAN, interpolate to 60fps with RIFE, and export as ProRes — hands-free.

🔌

Node-Based Control

Every parameter is exposed as a node input. Adjust CFG scale, seed, resolution, duration, and sampling steps individually. Connect parameter nodes to sliders for real-time tweaking. Branch your workflow to test multiple settings simultaneously and compare outputs side by side before committing to a full render.

🚀

Integration with Other Models

Use ControlNet for precise pose and depth control. Apply IP-Adapter for style transfer from reference images. Chain Seedance output into face restoration models or style-transfer networks. ComfyUI is the only interface where you can combine Seedance with the entire ecosystem of open-source AI models in a single graph.

PREREQUISITES

What You Need Before Starting

Gather these requirements before installing Seedance ComfyUI nodes. API-based workflows have minimal hardware requirements since inference runs on remote servers.

Software Requirements

Python 3.10 or 3.11 — Required by ComfyUI. Python 3.12+ may have compatibility issues with some nodes
Git — For cloning ComfyUI and custom node repositories
ComfyUI (latest) — The base application. We cover installation in the next section
Seedance API Key — From BytePlus, fal.ai, or Replicate
pip / conda — For managing Python dependencies

Hardware Requirements

GPU (API mode): Any GPU that runs ComfyUI — even integrated graphics work
GPU (local post-processing): 8GB+ VRAM recommended for upscaling and interpolation
RAM: 16GB minimum, 32GB recommended for complex workflows
Storage: 5GB for ComfyUI + nodes + cached outputs
Internet: Required for API calls. Stable connection recommended for large video downloads

No GPU at all? You can run ComfyUI in CPU-only mode for pure API workflows. Seedance inference happens on remote servers, so your local hardware only needs to handle the ComfyUI interface and any local post-processing nodes. For cloud GPU options, see our local setup guide.