Deployment

Choose the deployment model that fits your security and operational requirements.

Deployment Models

Model	Infrastructure	Best For
Cloud (SaaS)	Fully managed by Edison	Quick start, small teams.
Self-Hosted	Your infrastructure	Air-gapped, strict data residency.
Hybrid	Cloud dashboard + local MCP	Cloud convenience with on-premise data.

Self-Hosted Deployment (Docker)

Docker is the recommended method for self-hosting.

1. Prerequisites

Docker & Docker Compose
PostgreSQL 14+ (or SQLite for testing)
2GB RAM minimum (4GB recommended)

2. Configuration

Create a docker-compose.yml:

services:
  edison-watch:
    image: ghcr.io/edison-watch/edison-watch:latest
    ports:
      - "3000:3000"  # MCP gateway
      - "3001:3001"  # API server
    environment:
      - DATABASE_URL=postgresql://user:pass@db:5432/edison
      - SECRET_KEY=${SECRET_KEY}
    volumes:
      - ./data:/app/data
    restart: unless-stopped
 
  db:
    image: postgres:14
    environment:
      - POSTGRES_PASSWORD=pass
      - POSTGRES_DB=edison
    volumes:
      - pgdata:/var/lib/postgresql/data

3. Start

Generate a secure key and start:

export SECRET_KEY=$(openssl rand -hex 32)
docker-compose up -d

4. Verify

curl http://localhost:3001/health
# {"status": "healthy"}

Alternative: Python Install

uv pip install edison-watch
edison-watch serve --config config.json

Production Checklist

HTTPS: Use a reverse proxy (Nginx/Caddy) for TLS termination.
Database: Enable SSL and automated backups for PostgreSQL.
Monitoring: Scrape /metrics for Prometheus or use /health for uptime.
Logs: Default output is structured JSON for easy SIEM ingestion.

Need enterprise deployment assistance? Contact our team.

On this page