# Ollama - Run Large Language Models Locally
# Docs: https://ollama.ai

services:
  ollama:
    container_name: ollama
    image: ollama/ollama:latest
    restart: unless-stopped

    env_file:
      - .env

    volumes:
      - ./models:/root/.ollama

    networks:
      - homelab

    # GPU Support (NVIDIA GTX 1070)
    # Uncomment the deploy section below to enable GPU acceleration
    # Prerequisites:
    # 1. Install NVIDIA Container Toolkit on host
    # 2. Configure Docker to use nvidia runtime
    # deploy:
    #   resources:
    #     reservations:
    #       devices:
    #         - driver: nvidia
    #           count: 1
    #           capabilities: [gpu]

    labels:
      # Traefik (API only, no web UI)
      traefik.enable: true
      traefik.docker.network: homelab

      # API endpoint
      traefik.http.routers.ollama.rule: Host(`ollama.fig.systems`)
      traefik.http.routers.ollama.entrypoints: websecure
      traefik.http.routers.ollama.tls.certresolver: letsencrypt
      traefik.http.services.ollama.loadbalancer.server.port: 11434

      # SSO Protection for API
      traefik.http.routers.ollama.middlewares: tinyauth

      # Homarr Discovery
      homarr.name: Ollama (LLM)
      homarr.group: Services
      homarr.icon: mdi:brain

networks:
  homelab:
    external: true