Runs Blender headlessly to generate 3D depth maps and blockout/wireframe renders for storyboard shots.
Provides optional local vision-model evaluation of rendered storyboard images for scoring framing/composition and suggesting refinements.
Stores and retrieves successful storyboard recipes, including camera settings and prompts, in a local SQLite knowledge database for reuse.
Click on "Deploy Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type @ followed by the MCP server name and your instructions, e.g., "@StoryboardStudio MCP make a storyboard of a knight fighting a dragon at sunset"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
MCP-native pre-production system that generates multi-granularity storyboard assets from text prompts:
2D SVG Motion – timing-validated animation code
3D Depth / Blockout – Blender headless (bpy) Depth maps
Final Look (SDXL) – ControlNet Depth-guided image generation
LLM agents (Claude Desktop, Cursor, etc.) control the entire pipeline via Stdio + FastMCP.
Includes Vision-based self-correction and a SQLite Knowledge Database of successful recipes.
[ MCP Client ] (Claude Desktop / Cursor / Custom Agent)
│
▼ (Stdio / JSON-RPC)
[ StoryboardStudio MCP Server (Python FastMCP) ]
│
├─► SVG Motion Engine → .svg
├─► Blender CLI Runner → Depth / Wireframe PNG
├─► ComfyUI / SD-WebUI Client → Final Image (ControlNet)
├─► Vision Evaluator → Score + Refinement Delta
└─► Knowledge Engine (SQLite) → storyboard_knowledge.dbRelated MCP server: Eversince MCP Server
Component | Version / Notes |
Python | 3.11+ |
Blender | 4.0+ (with |
ComfyUI or A1111 | Running on |
Ollama (optional) | For local Vision evaluation |
GPU | RTX 3090+ recommended (target < 60 s/shot) |
pip install -r requirements.txtpython scripts/init_db.pypython server.pyAdd to your MCP config (claude_desktop_config.json or Cursor settings):
{
"mcpServers": {
"storyboard-studio": {
"command": "python",
"args": ["/absolute/path/to/storyboard-studio-mcp/server.py"],
"env": {
"BLENDER_PATH": "blender",
"COMFYUI_URL": "http://127.0.0.1:8188",
"OUTPUT_DIR": "/absolute/path/to/storyboard-studio-mcp/outputs"
}
}
}
}See mcp_config.json for a ready-to-use template.
Tool | Description |
| Save & validate SVG animation code |
| Run Blender headless → Depth PNG |
| ComfyUI/WebUI + ControlNet Depth → final image |
| Vision LLM scores framing / composition |
| Store high-rated camera + prompt to SQLite |
| Retrieve past successful recipes by tag |
storyboard-studio-mcp/
├── server.py # FastMCP entry point (all tools)
├── modules/
│ ├── blender_runner.py # Blender CLI + Depth injection
│ ├── sdxl_client.py # ComfyUI / A1111 API client
│ ├── vision_evaluator.py # Vision LLM evaluation
│ └── knowledge_db.py # SQLite recipe store
├── scripts/
│ └── init_db.py # Create storyboard_knowledge.db
├── outputs/
│ ├── svg/
│ ├── blender/
│ └── sdxl/
├── storage/
│ └── storyboard_knowledge.db
├── tests/
├── mcp_config.json
├── requirements.txt
└── README.mdPhase 1 (MVP) – Core pipeline
Phase 2 – Self-Correction Loop + Knowledge Engine
See the original Technical Specifications / PRD / MRD for full details.
MIT
StoryboardStudio MCP – Reduce storyboard iteration friction to near zero.
This server cannot be deployed
AI visual generation agent: multi-pipeline rendering, prompt crafting, and image composition.
Build and run visual creative-production workflows from your AI agent.
Generate and edit images, video, voice, lip-sync and 3D models from your AI agent.
On-brand creative studio for AI agents: images, video, audio, and 3D.
