User Prompt
    │
    ▼
┌─────────────────────┐
│    Planner Agent    │  ← Decomposes prompt into objects
└─────────┬───────────┘
          │
          ▼
┌─────────────────────┐
│  Image Gen Agent    │  ← DALL-E 3 / FLUX / Qwen (2D preview)
└─────────┬───────────┘
          │
    [User confirms preview]
          │
          ▼
┌─────────────────────┐
│  Generation Agent   │  ← Tripo3D API / Hunyuan3D (local)
└─────────┬───────────┘
          │
          ▼
┌─────────────────────────────────────┐
│  Intelligent PostProcessing Agent   │
│  ├── Overhang analysis (24 angles)  │
│  ├── Support strategy decision      │
│  └── Optimal orientation on bed     │
└─────────┬───────────────────────────┘
          │
          ▼
    Print-Ready STL

desk_organizer analysis:

✅ Printable without supports in recommended orientation.

Complexity: EASY

AI Analysis:
  - Geometry complexity: MEDIUM
  - Max overhang angle: 38.5°
  - Bed contact area: 1 250 mm²
  - No internal cavities detected
  - Recommended: rotate 180° around X axis

docker pull ghcr.io/teslaproduuction/3d-agent-mcp:latest

cp .env.example .env
# Fill in API keys

docker-compose up -d
# → http://localhost:7860

# Install UV
winget install --id=astral-sh.uv -e        # Windows
curl -LsSf https://astral.sh/uv/install.sh | sh  # Linux/macOS

# Clone and setup
git clone https://github.com/teslaproduuction/3d-agent-mcp.git
cd 3d-agent-mcp
uv venv --python 3.10
uv sync --all-extras

# Configure
cp .env.example .env
# Edit .env with your API keys

# Run
uv run python ui/gradio_app.py
# → http://localhost:7860

cp .env.example .env
# Edit .env

docker-compose up -d --build
# → http://localhost

python -m venv .venv
source .venv/bin/activate      # Linux/macOS
.venv\Scripts\activate         # Windows

pip install -r requirements.txt
cp .env.example .env
python ui/gradio_app.py        # → http://localhost:7860

{
  "mcpServers": {
    "3d-agent": {
      "command": "python",
      "args": ["/path/to/3d-agent-mcp/mcp_server/server.py"],
      "env": {
        "TRIPO_API_KEY": "your_key",
        "OPENAI_API_KEY": "your_key"
      }
    }
  }
}

User: Generate a phone stand for 3D printing

Claude: [calls generate_3d_model tool]
✅ Model generated and optimized for printing!
   - File: outputs/models/phone_stand_optimized.stl
   - Supports: none required
   - Orientation: base-down
   - Print time: ~2h 15min

default_settings:
  # Image generation
  image_generation:
    provider: "local"     # local | dalle3 | sdxl | flux

  # 3D generation
  generation:
    api_provider: "local" # local | tripo | meshy
    face_limit: 10000

  # Post-processing
  postprocessing:
    mode: "intelligent"   # AI decides automatically
    auto_orient: true
    max_overhang_angle: 45.0

  # Printer profile
  printer:
    build_volume: [220, 220, 250]  # mm — Ender 3 / Bambu A1
    nozzle_diameter: 0.4
    material: "PLA"

# LLM backend
llm:
  default_provider: "ollama"   # ollama | openai | anthropic
  local:
    ollama_models: ["qwen2.5:32b", "qwen2.5:7b"]

3d-agent-mcp/
├── agents/                              # AI agents
│   ├── coordinator.py                   # Pipeline orchestrator
│   ├── planner_agent.py                 # Scene decomposition
│   ├── image_generation_agent.py        # 2D preview
│   ├── generation_agent.py              # 3D API calls
│   └── intelligent_postprocessing_agent.py
│
├── api_clients/                         # API wrappers
│   ├── llm_client.py                    # OpenAI / Anthropic / Ollama
│   ├── image_api_client.py              # DALL-E / SDXL / FLUX
│   └── tripo_client.py                  # Tripo3D
│
├── mcp_server/                          # MCP server
│   ├── server.py                        # Tool definitions
│   └── README.md                        # MCP API docs
│
├── ui/                                  # Gradio web UI
│   ├── gradio_app.py                    # Main app
│   └── tabs/, handlers/, components/
│
├── postprocessing/                      # Geometry analysis
├── docker/                              # Local model containers
│   ├── hunyuan3d/                       # Hunyuan3D-2 (local 3D)
│   ├── flux/                            # FLUX.1 (local image gen)
│   ├── comfyui/                         # ComfyUI
│   └── nginx/                           # Reverse proxy
│
├── tests/
├── config.yaml                          # Main config
├── .env.example                         # API key template
├── docker-compose.yml                   # Full Docker stack
└── pyproject.toml

# Run tests
pytest tests/

# Format
black .

# Lint
flake8 .

# Type check
mypy .