YOLO MCP Server
by GongRzhe
Computer vision with YOLO object detection through AI assistants
ai-ml Python Advanced Self-hostable No API key
⭐ 30 stars 📅 Updated: 1y ago
Description
An MCP server integrating YOLO with AI assistants for computer vision tasks. Provides 13 tools covering object detection, image classification, segmentation, real-time camera detection, and model management (training, validation, export). Supports multiple YOLO models, configurable confidence thresholds, and comprehensive image analysis combining detection, classification, and segmentation in one pass.
✅ Best for
Developers needing full YOLO-based computer vision capabilities through AI assistants
⏭️ Skip if
You only need basic image processing or prefer a cloud-based vision API
💡 Use cases
- Detecting and classifying objects in images via AI commands
- Real-time object detection using webcam feeds
- Training custom YOLO models on specific datasets
- Segmenting objects with boundary masks for analysis
- Exporting trained models to ONNX and other formats
👍 Pros
- ✓ Rich 13-tool set covering detection, segmentation, classification, and training
- ✓ Real-time camera integration for live detection
- ✓ Model lifecycle management (train, validate, export)
- ✓ Comprehensive analysis combining multiple CV tasks in one call
👎 Cons
- ✗ Heavy dependencies (ultralytics, opencv)
- ✗ YOLO model files can be large (hundreds of MB)
- ✗ Manual installation via setup.py — no pip package published
🔧 Exposed tools (11 tools)
| Tool | Category | Description |
|---|---|---|
| start_camera_detection | camera | Start real-time camera-based object detection |
| get_camera_detections | camera | Retrieve latest camera analysis results |
| stop_camera_detection | camera | Stop camera detection operations |
| classify_image | classification | Categorize entire image content with top-k results |
| analyze_image_from_path | detection | Detect objects in images with configurable confidence thresholds |
| comprehensive_image_analysis | detection | Combined detection, classification, and segmentation in one pass |
| list_available_models | model-management | Display accessible YOLO models on the system |
| train_model | model-management | Create custom detection models from datasets |
| validate_model | model-management | Evaluate model performance metrics |
| export_model | model-management | Convert models to ONNX and other formats |
| segment_objects | segmentation | Identify object boundaries and create masks |
💡 Tips & tricks
Run python setup.py first to download YOLO models automatically. Use comprehensive_image_analysis for a one-call combination of detection, classification, and segmentation.
Quick info
- Author
- GongRzhe
- License
- MIT
- Runtime
- Python 3.10+
- Transport
- stdio
- Category
- ai-ml
- Difficulty
- Advanced
- Self-hostable
- ✅
- Auth
- —
- Docker
- —
- Version
- 1.0.0
- Updated
- Mar 11, 2025
Client compatibility
- ❓ Claude Code
- ❓ Cursor
- ❓ VS Code Copilot
- ❓ Gemini CLI
- ❓ Windsurf
- ❓ Cline
- ❓ JetBrains AI
- ❓ Warp
Platforms
🍎 macOS 🐧 Linux 🪟 Windows