Ultralytics MCP Server

services: # RCT Detector Platform - Main Application rct-detector: build: context: . dockerfile: Dockerfile.ultralytics container_name: rct-detector-platform volumes: # Workspace for trained models - trained_models:/workspace/trained_models # Built-in datasets (mount your datasets folder here) - ../datasets:/ultralytics/YOLO_MultiLevel_Datasets:ro # Custom datasets volume - custom_datasets:/ultralytics/custom_datasets environment: - NVIDIA_VISIBLE_DEVICES=all - NVIDIA_DRIVER_CAPABILITIES=compute,utility - CUDA_VISIBLE_DEVICES=0 runtime: nvidia gpus: all deploy: resources: reservations: devices: - driver: nvidia count: all capabilities: [gpu] ports: - "8501:8501" # Streamlit Web Interface - "8888:8888" # Jupyter (optional) command: > bash -c " mkdir -p /workspace/trained_models && mkdir -p /ultralytics/custom_datasets && cd /ultralytics && streamlit run main_dashboard.py --server.address 0.0.0.0 --server.port 8501 --server.headless true --server.enableCORS false --server.enableXsrfProtection false --server.maxUploadSize 10000 " stdin_open: true tty: true shm_size: '8gb' ulimits: memlock: -1 stack: 67108864 networks: - rct-network restart: unless-stopped # MCP Server for N8N Integration (Optional) mcp-server: build: context: . dockerfile: Dockerfile.mcp-connector container_name: rct-mcp-server ports: - "8092:8092" # MCP Server endpoint environment: - NODE_ENV=production - ULTRALYTICS_CONTAINER=rct-detector-platform networks: - rct-network restart: unless-stopped depends_on: - rct-detector volumes: trained_models: driver: local custom_datasets: driver: local networks: rct-network: driver: bridge