services:
  gateway:
    image: caddy:2.8-alpine
    restart: unless-stopped
    depends_on:
      - llama-server
    env_file:
      - .env
    ports:
      - "80:80"
      - "443:443"
    volumes:
      - ./Caddyfile:/etc/caddy/Caddyfile:ro
      - ./site:/srv/site:ro
      - caddy_data:/data
      - caddy_config:/config

  llama-server:
    image: ghcr.io/ggml-org/llama.cpp:server
    restart: unless-stopped
    env_file:
      - .env
    environment:
      HF_TOKEN: ${HF_TOKEN:-}
    command:
      - --hf-repo
      - ${LLAMA_HF_REPO}
      - --hf-file
      - ${LLAMA_HF_FILE}
      - --host
      - 0.0.0.0
      - --port
      - "8000"
      - --ctx-size
      - ${LLAMA_CTX_SIZE}
      - --threads
      - ${LLAMA_THREADS}
      - --parallel
      - ${LLAMA_PARALLEL}
    expose:
      - "8000"
    volumes:
      - llama_cache:/root/.cache/llama.cpp

  agentbox:
    build:
      context: .
      dockerfile: Dockerfile.agentbox
    restart: unless-stopped
    env_file:
      - .env
    environment:
      TZ: ${TZ}
    tty: true
    stdin_open: true
    working_dir: /srv/workspaces
    volumes:
      - ./workspaces:/srv/workspaces

volumes:
  caddy_data:
  caddy_config:
  llama_cache: