seedance

# Seedance AI video generation using ByteDance's Seedance model via fal.ai API. Supports text-to-video and image-to-video generation with audio. ## Customization **Before executing, check for user customizations at:** `~/.claude/skills/PAI/USER/SKILLCUSTOMIZATIONS/Seedance/` ## Voice Notification **When executing a workflow, do BOTH:** 1. **Send voice notification**: ```bash curl -s -X POST http://localhost:8888/notify \ -H "Content-Type: application/json" \ -d '{"message": "Running the WORKFLOWNAME workflow in the Seedance skill"}' \ > /dev/null 2>&1 & ``` 2. **Output text notification**: ``` Running the **WorkflowName** workflow in the **Seedance** skill... ``` ## Workflow Routing | Workflow | Trigger | File | |----------|---------|------| | **Generate** | "generate video", "create video", "text to video", "image to video", "animate this" | `Workflows/Generate.md` | ## Quick Reference ```bash # Text-to-video bun run ~/.claude/skills/Seedance/Tools/Generate.ts \ --prompt "A cat walking through a neon city" \ --duration 5 --resolution 720p --output ~/Downloads/video.mp4 # Image-to-video (animate a still image) bun run ~/.claude/skills/Seedance/Tools/Generate.ts \ --image ~/Downloads/photo.png \ --prompt "Camera slowly zooms in, gentle wind" \ --duration 5 --output ~/Downloads/animated.mp4 ``` **Environment:** Requires `FAL_KEY` in `~/.claude/.env` **Full API details:** `ApiReference.md` ## Current Model **Seedance 1.5 Pro** (fal.ai) — Latest available via API. Seedance 2.0 is China-only (Dreamina) as of Feb 2026. When it hits fal.ai, update the endpoint constant in `Tools/Generate.ts`. ## Examples **Example 1: Generate video from text** ``` User: "generate a video of ocean waves at sunset" -> Invokes Generate workflow -> Calls fal.ai Seedance text-to-video -> Downloads MP4 to ~/Downloads/ -> Opens video for preview ``` **Example 2: Animate a still image** ``` User: "animate this image" (with reference to an image) -> Invokes Generate workflow (image-to-video mode) -> Uploads image + prompt to fal.ai -> Returns animated video ``` **Example 3: High-res cinematic clip** ``` User: "generate a cinematic 1080p video of a samurai walking through fog, 10 seconds" -> Invokes Generate workflow -> --resolution 1080p --duration 10 -> Full HD output with audio ```