DEV Community

Evan-dong
Evan-dong

Posted on

How to Use Seedance 2.0 API: Three Integration Paths for AI Video Generation

#ai #video #api #tutorial

If you need programmatic access to ByteDance's Seedance 2.0 — the multimodal AI video model that supports @-references, V2V editing, and frame-accurate audio — this guide walks through three practical integration paths: a no-code playground, an agent skill, and direct API calls.

This covers setup, all three generation modes, pricing math, and the tips I wish I'd known earlier.

What Seedance 2.0 Actually Supports

Before jumping into integration, here's what makes this model worth the effort:

  • Multimodal @-reference system: Up to 9 images + 3 videos + 3 audio tracks in a single generation request
  • Video-to-video editing: Modify specific elements in existing video while preserving structure
  • Frame-accurate audio sync: Auto-generated dialogue, SFX, and BGM matching every frame
  • Multi-shot narratives: Structured sequences with camera cuts and consistent character identity
  • Pay-as-you-go pricing: No subscription — credit-based billing

Seedance 2.0 generated scene — cinematic interior with volumetric lighting

Path 1: Web Playground (No Code Required)

Best for: testing prompts, evaluating quality, understanding model behavior before committing to integration.

  1. Sign up at evolink.ai
  2. Navigate to Playground → Seedance 2.0
  3. Configure parameters (model, prompt, duration, resolution, aspect ratio)
  4. Click Generate

The playground exposes all three generation modes with a visual interface and cost calculator. Good for building intuition before writing code.

Path 2: ClawHub Skill (Fastest for Agent Users)

If you use OpenClaw or Claude Code, this is the quickest path to generation.

Install:

  1. Visit ClawHub: seedance-2-video-gen
  2. Click "Install Skill"
  3. Set your EVOLINK_API_KEY environment variable
  4. Describe what you want — the skill handles parameters, polling, and delivery

Example conversation:

You: Generate a 5-second video of a glass frog with a beating heart
Skill: Starting your video now — this usually takes 1-3 minutes.
       ✅ Done! Here's your video: [URL]
Enter fullscreen mode Exit fullscreen mode

Best for: rapid prototyping, creative exploration, non-technical users in the agent ecosystem.

Path 3: Direct API Integration (Production-Ready)

For applications, batch processing, and custom workflows.

Step 1: Get your API key 

export EVOLINK_API_KEY="your_key_here"
Enter fullscreen mode Exit fullscreen mode

Step 2: Submit a generation task 

curl --request POST \
  --url https://api.evolink.ai/v1/videos/generations \
  --header 'Authorization: Bearer YOUR_API_KEY' \
  --header 'Content-Type: application/json' \
  --data '{
    "model": "seedance-2.0-text-to-video",
    "prompt": "A macro lens focuses on a green glass frog on a leaf. The focus gradually shifts from its smooth skin to its completely transparent abdomen, where a bright red heart is beating powerfully and rhythmically.",
    "duration": 8,
    "quality": "720p",
    "aspect_ratio": "16:9",
    "generate_audio": true
  }'
Enter fullscreen mode Exit fullscreen mode

Response:

{
  "id": "task_abc123",
  "status": "processing",
  "estimated_time": 90
}
Enter fullscreen mode Exit fullscreen mode

Step 3: Poll for results 

curl --request GET \
  --url https://api.evolink.ai/v1/tasks/task_abc123 \
  --header 'Authorization: Bearer YOUR_API_KEY'
Enter fullscreen mode Exit fullscreen mode

Completed response:

{
  "id": "task_abc123",
  "status": "completed",
  "video_url": "https://cdn.evolink.ai/videos/...",
  "duration": 8,
  "cost": 80
}
Enter fullscreen mode Exit fullscreen mode

Or skip polling entirely by passing callback_url in your initial request.

The Three Generation Modes

Text-to-Video

Prompt-only generation. No reference assets needed.

{
  "model": "seedance-2.0-text-to-video",
  "prompt": "Cinematic aerial shot of a futuristic city at sunrise, soft clouds, reflective skyscrapers, smooth camera movement",
  "duration": 5,
  "quality": "720p"
}
Enter fullscreen mode Exit fullscreen mode

Best for: concept visualization, trend content, creative exploration.

Image-to-Video

Animates still images. One image = first-frame animation. Two images = first-to-last-frame transition.

{
  "model": "seedance-2.0-image-to-video",
  "prompt": "Camera slowly pushes in, the still scene comes to life",
  "image_urls": ["https://example.com/product.jpg"],
  "duration": 5,
  "aspect_ratio": "adaptive"
}
Enter fullscreen mode Exit fullscreen mode

Best for: product demos, social media content, photo animation.

Reference-to-Video

Maximum control. Accepts images, video clips, and audio as simultaneous references.

{
  "model": "seedance-2.0-reference-to-video",
  "prompt": "Use video 1's camera movement with audio 1 as background music",
  "image_urls": ["https://example.com/character.jpg"],
  "video_urls": ["https://example.com/motion-ref.mp4"],
  "audio_urls": ["https://example.com/bgm.mp3"],
  "duration": 10,
  "quality": "720p"
}
Enter fullscreen mode Exit fullscreen mode

Best for: advanced editing, style transfer, multimodal composition, video extension.

Pricing Math

Credit-based, no subscription. 1 credit = $0.01 USD.

Text-to-video & Image-to-video:

Resolution Credits/second
480p 4.63
720p 10.00

Reference-to-video: (input duration + output duration) × resolution rate

Real-world cost examples:

Scenario Calculation Monthly Cost
Short-form creator (10 vids/day, 5s, 720p) 10 × 5 × 10 × 30 $150
Product team (20 demos/week, 8s, 720p) 20 × 8 × 10 × 4 $64

Which Path Should You Choose?

Method Best For Setup Time Technical Skill
Playground Testing, evaluation 1 minute None
ClawHub Skill Rapid prototyping, creative work 2 minutes None
Direct API Production apps, automation 15 minutes Developer-level

Start with Playground to understand behavior, use ClawHub Skill for daily creative work, integrate the API when you're ready for production.

Three Tips That Save Time and Credits

1. Use aspect_ratio: "adaptive" for irregular images — lets the model choose the best fit instead of cropping.

2. Set duration: -1 for smart duration — the model determines optimal length based on content. You're charged for actual output, not maximum.

3. Keep reference videos short — input video duration counts toward cost in reference-to-video mode. Trim references to 5-10 seconds:

ffmpeg -i long-video.mp4 -t 5 -c copy motion-ref.mp4
Enter fullscreen mode Exit fullscreen mode

Getting Started Checklist

Week 1: Test all three modes in Playground. Collect reference materials (character designs, motion templates, style references).

Week 2: Choose your integration path. Set up API or install ClawHub Skill. Implement error handling and retry logic.

Week 3+: Start with simple text-to-video, gradually add multimodal references, monitor costs and success rates, build prompt templates.

I documented the full API reference and code examples here: Seedance 2.0 API on EvoLink

Top comments (0)