Quick Start

Get AutoFlow running locally in under 5 minutes.

Prerequisites

• Node.js 18 or later
• npm 9+
• Docker (optional — for containerized deployment)

1. Clone and install

git clone https://github.com/autoflow-hq/autoflow.git
cd autoflow
npm install

2. Start the API server

The development server starts on port 3000 with hot-reload.

npm run dev

Verify it started:

curl http://localhost:3000/health

# Response:
{
  "status": "ok",
  "templates": 3,
  "runs": { "total": 0, "running": 0, "completed": 0, "failed": 0 }
}

3. Run your first workflow

Start the Lead Enrichment workflow with a sample lead:

curl -X POST http://localhost:3000/api/runs \
  -H "Content-Type: application/json" \
  -d '{
    "templateId": "tpl-lead-enrich",
    "input": {
      "name": "Jane Smith",
      "email": "jane@acme.com",
      "company": "Acme Corp"
    },
    "config": {
      "crmIntegration": "hubspot",
      "idealCustomerProfile": "Series A SaaS, 50-200 employees, US-based"
    }
  }'

The API returns immediately with a run object (status pending → running → completed). Poll the run ID to track progress:

curl http://localhost:3000/api/runs/<run-id>

4. Explore templates

List all available templates:

curl http://localhost:3000/api/templates

Get a template's full definition including config fields and step definitions:

curl http://localhost:3000/api/templates/tpl-lead-enrich

Get sample input/output to understand what a template expects:

curl http://localhost:3000/api/templates/tpl-lead-enrich/sample

5. Run via Docker

The included docker-compose.yml starts the API backend and React dashboard:

docker compose up

• API: http://localhost:8000
• Dashboard: http://localhost:3001

Production deployment

AutoFlow backend runs on Azure; the dashboard is hosted on Vercel. See infra/README.md for full setup instructions including DNS, TLS, and CI/CD configuration.