version: "3.8"

services:
  ollama:
    build:
      context: .
      dockerfile: Dockerfile
    container_name: ollama
    restart: unless-stopped
    ports:
      - "11434:11434"
    volumes:
      - ollama_data:/root/.ollama
    deploy:
      resources:
        reservations:
          devices:
            - driver: nvidia
              count: all
              capabilities: [gpu]
    environment:
      - OLLAMA_HOST=0.0.0.0
      - OLLAMA_KEEP_ALIVE=24h

  litellm:
    image: ghcr.io/berriai/litellm:main-latest
    container_name: litellm
    restart: unless-stopped
    depends_on:
      - ollama
    ports:
      - "4000:4000"
    volumes:
      - ./litellm-config.yaml:/app/config.yaml
    command: ["--config", "/app/config.yaml", "--port", "4000"]

  claude-code:
    image: node:20-slim
    container_name: claude-code
    restart: unless-stopped
    depends_on:
      - litellm
    environment:
      # Point Claude Code at LiteLLM which fully speaks the Anthropic API
      - ANTHROPIC_BASE_URL=http://litellm:4000
      - ANTHROPIC_API_KEY=ollama
    volumes:
      - workspace:/workspace
      - claude_config:/root/.claude
    working_dir: /workspace
    command: >
      sh -c "npm install -g @anthropic-ai/claude-code &&
             mkdir -p /root/.claude &&
             echo '{\"model\": \"qwen2.5-coder:32b\"}' > /root/.claude/settings.json &&
             echo 'Claude Code ready. Run: claude' &&
             tail -f /dev/null"
    stdin_open: true
    tty: true

volumes:
  ollama_data:
    driver: local
  workspace:
    driver: local
  claude_config:
    driver: local