Skip to content

Docker Installation

Universal Docker setup that works on any platform with Docker support.

⚠️ Important Limitations

  • macOS: Docker does not support GPU acceleration. For 10x better performance, use macOS native setup
  • Linux: Limited GPU support. For full NVIDIA GPU acceleration, use Linux native setup
  • Ollama on macOS: Can be memory-intensive without GPU acceleration

Prerequisites

  • Docker and Docker Compose installed
  • At least 8GB RAM available for Docker
  • 10GB free disk space

Quick Start

  1. Start the services:
docker compose -f docker/docker-compose.yml up --build
  1. Check if services are running:
docker compose -f docker/docker-compose.yml logs

You should see logs from all services, with Ollama downloading the gemma3:4b model.

  1. Install agent-cli:
uv tools install agent-cli
# or: pip install agent-cli
  1. Test the setup: 
    agent-cli autocorrect "this has an eror"

Services Overview

The Docker setup provides:

Service Image Port Purpose
ollama Custom build 11434 LLM server with gemma3:4b
whisper rhasspy/wyoming-whisper 10300 Speech-to-text (large-v3)
piper rhasspy/wyoming-piper 10200 Text-to-speech (ryan-high)
openwakeword rhasspy/wyoming-openwakeword 10400 Wake word detection

Configuration Files

The Docker setup uses:

  • docker/docker-compose.yml - Service orchestration
  • docker/Dockerfile - Custom Ollama container
  • Data volumes for model persistence

Important Limitations

  • macOS: No GPU acceleration (10x slower than native)
  • Linux: Limited GPU support
  • Memory: Requires 8GB+ RAM for smooth operation

Managing Services

# Start services in background
docker compose -f docker/docker-compose.yml up -d

# Stop services
docker compose -f docker/docker-compose.yml down

# View logs
docker compose -f docker/docker-compose.yml logs -f

# Restart a specific service
docker compose -f docker/docker-compose.yml restart ollama

Data Persistence

Services store data in local directories:

  • ./ollama/ - Ollama models and config
  • ./whisper-data/ - Whisper models
  • ./piper-data/ - Piper voice models
  • ./openwakeword-data/ - Wake word models

Troubleshooting

Common Issues

  • Slow performance: Use native setup for better performance
  • Memory issues: Increase Docker memory allocation to 8GB+
  • Port conflicts: Change port mappings in docker/docker-compose.yml

Alternative: Native Installation

For better performance, consider platform-specific native installation: