Quick start
Install PyOrchestrator with Docker Compose in minutes: clone, configure .env, start the stack, and sign in to the control plane.
Requirements
- Docker 24+ and Docker Compose v2
- 4 GB RAM (recommended 8 GB with observability stack)
- Linux or macOS (cgroups for runtime - optional)
Installation
git clone https://github.com/PyOrchestrator/PyOrchestrator.git
cd PyOrchestrator
git checkout v0.1.13 # or the latest tag: git describe --tags --abbrev=0
cp .env.example .env
docker compose up --build
The first launch takes a few minutes: building images, initializing PostgreSQL, MinIO, seed demo scripts.
Current version: v0.1.13 - release notes.
Login
| Field | Default |
|---|---|
admin@pyorchestrator.local |
|
| Password | admin |
After logging in, open Scripts - demo objects will be loaded (Weather, Crypto, DAP API Poster, etc.).
First run of the script
- Open any script → Editor
- Click Run - live output will appear below
- Stop — stop sandbox via Redis control channel
Services and ports
| Service | Port | Destination |
|---|---|---|
| frontend | 5173 | React UI (dev) |
| backend | 8000 | REST API |
| postgres | 5432 | Metadata |
| redis | 6379 | Queue jobs |
| minio | 9000 (+ 9001 with MINIO_CONSOLE_ENABLED=true) |
S3 API / optional Console |
| grafana | 3000 | Dashboards |
| prometheus | 9090 | Metrics |
| mcp | 8010 | MCP HTTP |
Ports are configured in .env - see Configuration.
Production
docker compose -f docker-compose.yml -f docker-compose.prod.yml up -d --build
More details: Deployment.
MCP for Cursor
cd mcp && pip install -e .
Add the config from mcp/cursor-mcp.example.json to Cursor Settings → MCP.
More details: MCP.