Installation¶
This guide covers different ways to install and deploy Marionette.
Prerequisites¶
- Go 1.22+ (for building from source)
- PostgreSQL 15+ (database)
- Docker (recommended for local development)
Quick Install with Docker Compose¶
The fastest way to get started:
# Clone the repository
git clone https://github.com/chunlea/marionette.git
cd marionette
# Start all services
docker compose up -d
This starts:
- Server on ports 8080 (API), 8081 (Admin), 9090 (gRPC)
- PostgreSQL on port 5432
- Agent ready to accept tasks
Building from Source¶
1. Clone and Install Dependencies¶
2. Setup Database¶
3. Build Binaries¶
This creates three binaries in ./bin/:
| Binary | Description |
|---|---|
server |
Main server (API, Admin, gRPC) |
agent |
Runner agent |
mctl |
CLI tool |
4. Start the Server¶
Kubernetes Deployment¶
Using Kustomize¶
# Development
kubectl apply -k deploy/kubernetes/overlays/dev
# Production
kubectl apply -k deploy/kubernetes/overlays/prod
Using Helm¶
# Add the chart repository (if hosted)
helm repo add marionette https://chunlea.github.io/marionette/charts
# Install
helm install marionette marionette/marionette \
--set postgresql.enabled=true \
--set server.replicas=3
Or install from local chart:
Configuration¶
After installation, configure Marionette using environment variables or a config file:
See Configuration for detailed options.
Verify Installation¶
# Check server health
curl http://localhost:8081/health/ready
# List sessions (should be empty)
./bin/mctl sessions list
# Check metrics
curl http://localhost:9091/metrics | head -20
Next Steps¶
- Quick Start - Run your first AI coding agent
- Configuration - Configure providers and storage
- CLI Reference - Learn the CLI commands