stack/voice/docker-compose.yml

version: '3.8'

services:
  voice-api:
    build: .
    ports:
      - "8000:8000"
    volumes:
      - ./voice_models:/app/voice_models
      - ./audio_files:/app/audio_files
    environment:
      - MODEL_PATH=/app/models/coqui_xtts
      - VOICE_CACHE_DIR=/app/voice_cache
      - WORKERS=4
    deploy:
      resources:
        limits:
          cpus: '2.0'
          memory: 4G
        reservations:
          cpus: '1.0'
          memory: 2G
    restart: unless-stopped

  tts-model:
    image: coqui/tts:latest
    ports:
      - "9000:9000"
    volumes:
      - ./models:/models
      - ./tts_cache:/tts_cache
    environment:
      - MODEL_NAME=x TTS
      - MODEL_PATH=/models/coqui_xtts
      - CACHE_DIR=/tts_cache
      - GPU_SUPPORT=${GPU_SUPPORT:-false}
    deploy:
      resources:
        limits:
          cpus: '4.0'
          memory: 8G
          ${GPU_LIMITS}
        reservations:
          cpus: '2.0'
          memory: 4G
    restart: unless-stopped

  redis:
    image: redis:alpine
    ports:
      - "6379:6379"
    volumes:
      - ./redis_data:/data
    command: redis-server --appendonly yes
    deploy:
      resources:
        limits:
          cpus: '0.5'
          memory: 256M
        reservations:
          cpus: '0.25'
          memory: 128M
    restart: unless-stopped

  # Optional: Speech-to-text service for voice input
  stt-service:
    image: vosk/kaldi:latest
    ports:
      - "9001:9001"
    volumes:
      - ./models/vosk:/models/vosk
    environment:
      - MODEL_PATH=/models/vosk/model
    deploy:
      resources:
        limits:
          cpus: '2.0'
          memory: 4G
        reservations:
          cpus: '1.0'
          memory: 2G
    restart: unless-stopped

volumes:
  voice_models:
    driver: local
  audio_files:
    driver: local
  models:
    driver: local
  tts_cache:
    driver: local
  redis_data:
    driver: local
  vosk_models:
    driver: local

networks:
  default:
    driver: bridge

# Environment variables for GPU support
# Set GPU_SUPPORT=true and provide GPU_LIMITS when using GPU
# Example: GPU_LIMITS=nvidia.com/gpu=1