Extensions/Seedance 2.0 (Sjinn.ai)
ComfyUI Extension

Seedance 2.0 (Sjinn.ai)

Run ByteDance Seedance 2.0 video generation in ComfyUI via Sjinn.ai API. Supports up to 6 image, 3 video and 2 audio references. Outputs native VIDEO plus first frame, last frame and frame batch.

By steptonite·Created 4 months ago·Updated 4 months ago· 2
steptonite/seedance2-sjinn-comfyui
Nodes1
On cloudLocal install
CategorySjinn.ai
Stars2
Updated4 months ago

Nodes (1)

🎬 Seedance 2.0 (Sjinn.ai)

Seedance 2.0 in ComfyUI without the VRAM bill — this node spends credits, not GPUs

Sjinn.ai
Readme

License ComfyUI

🎬 Seedance 2.0 ComfyUI Node — Sjinn.ai API

Run ByteDance Seedance 2.0 video generation directly in ComfyUI via the Sjinn.ai API.

Connect reference images, videos and audio clips, write a prompt, click Run. The node handles uploads, task creation and polling — you get back a native VIDEO output plus frame previews.

Workflow Preview

✨ Features

| | | |--|--| | 🖼️ Reference images | Up to 6 — connect any Load Image node | | 🎥 Reference videos | Up to 3 — connect vanilla Load Video node (≤720P, ≥2s) | | 🎵 Reference audio | Up to 2 — connect Load Audio node | | 📐 Aspect ratios | 16:9 · 9:16 · 1:1 · 4:3 · 3:4 | | ⏱️ Duration | 4–15 seconds | | ⚡ Speed mode | pro = quality · fast = speed | | 🎲 Seed | Fixed or random | | 🛡️ Smart rescue | Auto-rewrites prompt if content policy blocks it | | 📊 Progress bar | Real ComfyUI progress during uploads + generation | | 🎞️ Frame outputs | first_frame, last_frame, frames (IMAGE batch for VHS nodes) |

📦 Requirements

  • ComfyUI v0.14+ (native VIDEO and AUDIO types)
  • Sjinn.ai Pro+ account — https://sjinn.ai
  • opencv-python — auto-installed via requirements.txt

🚀 Installation

Via ComfyUI Manager (recommended):

  1. Manager → Install via Git URL
  2. Paste: https://github.com/steptonite/seedance2-sjinn-comfyui
  3. Install → restart ComfyUI

Manual:

cd ComfyUI/custom_nodes/
git clone https://github.com/steptonite/seedance2-sjinn-comfyui

# Mac / Linux
../../python/bin/python -m pip install -r seedance2-sjinn-comfyui/requirements.txt

# Windows portable
..\..\python_embeded\python.exe -m pip install -r seedance2-sjinn-comfyui/requirements.txt

Restart ComfyUI.

🔑 Credentials

Two credentials from sjinn.ai are required.

API Key

Profile → API Keys → Create → copy the key → paste into api_key field.

Session Token

  1. Open https://sjinn.ai/tools/seedance20-video (must be logged in, Pro+)
  2. F12 → Application → Cookies → https://sjinn.ai
  3. Find cookie: __Secure-next-auth.session-token
  4. Click the row → copy the full value from the bottom panel (starts with eyJ, ~500 chars)
  5. Paste into session_token field

⚠️ Token expires when your browser session ends. On 401 errors — grab a fresh one.

🕹️ How to Use

Find 🎬 Seedance 2.0 (Sjinn.ai) in the Sjinn.ai category (or search "Seedance").

  1. Connect images → image_1 … image_6
  2. Connect videos → video_1 … video_3 (must be ≤720P and ≥2 seconds)
  3. Connect audio → audio_1 / audio_2
  4. Write prompt — reference inputs with @Image1, @Image2, @Video1: 
    @Image1 is the character. @Video1 is the camera motion reference.
Cinematic slow motion, natural lighting.
  5. Set ratio, duration, speed_mode, seed
  6. Enter session_token and api_key
  7. Queue Prompt — progress bar shows upload → generation → done

Outputs: video (VIDEO) · first_frame · last_frame · frames (IMAGE batch)

📋 Parameters

| Parameter | Default | Description | |-----------|---------|-------------| | session_token | — | Browser cookie from sjinn.ai | | api_key | — | Sjinn.ai API key (for polling) | | prompt | — | Generation prompt | | image_1..6 | — | Reference images | | video_1..3 | — | Reference videos (≤720P, ≥2s) | | audio_1..2 | — | Reference audio clips | | ratio | 16:9 | Aspect ratio | | duration | 5 | Output length in seconds (4–15) | | speed_mode | pro | pro quality / fast speed | | accelerate_real_face | True | Preserve face consistency across frames | | smart_rescue | True | Auto-fix prompt on content policy block | | seed | -1 | -1 = random · fixed int = reproducible |

💰 Cost

100 credits per second of output. A 5-second video = 500 credits. Credits are refunded if the task fails content policy.

💡 Tips

  • Videos must be ≤720P — resize upstream if needed
  • Videos must be ≥2 seconds
  • speed_mode: fast for iteration, pro for finals
  • smart_rescue: off if you want the prompt used verbatim
  • Use @Image1, @Video1 etc. to explicitly assign roles in the prompt

🔐 Local Credentials File (optional)

Create sjinn_config.py in the node folder to pre-fill fields:

SJINN_API_KEY = "your-key"
SJINN_SESSION_TOKEN = "your-token"

Already in .gitignore — never commit this file.

🛠️ Troubleshooting

| Issue | Fix | |-------|-----| | 401 on upload/task | Session token expired — grab fresh cookie | | "Service temporarily unavailable" | Sjinn.ai server load — retry | | Task failed (status=2) | Enable smart_rescue or soften prompt | | Video too short error | Reference video must be ≥2 seconds | | video_1 / audio_1 socket missing | Update ComfyUI to v0.14+ | | first_frame is black | pip install opencv-python |

🇺🇦 Community

Ukrainian ComfyUI community — questions, workflows, tips: t.me/comfyui_ua

📄 License

MIT — see LICENSE