n8n

n8n is a fair-code workflow automation platform designed for technical teams who need the flexibility of custom code without abandoning a visual builder. Unlike SaaS-only tools, it can be fully self-hosted, making it suitable for manufacturers and IT/OT integrators who must keep data on-premise.

Key features

• Visual node editor with 1,544 pre-built integration nodes covering databases, APIs, messaging, and AI models
• Code steps anywhere — embed JavaScript or Python in any workflow node without switching tools
• MQTT trigger node — subscribe to MQTT broker topics directly, enabling IIoT event-driven workflows
• HTTP Request node — connect to any REST or GraphQL API, including MES, ERP, and SCADA systems with web interfaces
• AI agent orchestration — built-in LLM agent nodes (OpenAI, Anthropic, Gemini), vector store connections, and MCP (Model Context Protocol) client/server support
• Queue mode — horizontal scaling across multiple worker instances via Redis-backed queue
• Git version control — workflow definitions stored as JSON, diffable and branch-manageable (Business/Enterprise)
• Audit logging and SIEM streaming — log workflow executions to Datadog, Splunk, or other SIEM tools (Enterprise)

Manufacturing and IIoT use cases

n8n sits in the integration layer between OT and IT systems. Common patterns in industrial environments:

• IT/OT data routing — receive MQTT events from shop floor devices → enrich with ERP data → write to time-series database or data lake
• Alert escalation — threshold breach from historian/SCADA → webhook trigger → create ticket in Jira/ServiceNow → notify on-call via Slack/Teams
• Paperwork automation — production order completion event → auto-generate PDF report → email to quality team
• Security operations (SOAR) — Vodafone used n8n for threat intelligence enrichment, saving £2.2M annually
• Predictive maintenance data prep — poll sensor REST APIs on schedule → transform and normalize → push to ML pipeline

Deployment

n8n runs as a single Docker container for small teams or a scaled queue-mode cluster for enterprise:

docker run -it --rm \
  -p 5678:5678 \
  -v ~/.n8n:/home/node/.n8n \
  n8nio/n8n

For production: Docker Compose with PostgreSQL backend, Redis for queue mode, reverse proxy (Traefik/Nginx) for TLS. Helm chart available for Kubernetes.

Licensing

n8n uses a two-part fair-code license: the Sustainable Use License covers the core (self-hosting permitted, commercial resale of the platform restricted) and the n8n Enterprise License covers enterprise features. This is not an OSI-approved open-source license — vendors building managed n8n services must negotiate a commercial agreement. Community edition features are free forever.

Limitations

• Fair-code license restricts building competing managed-service offerings on top of n8n without a commercial agreement — not truly open-source by OSI definition
• No native OPC-UA or Modbus node — industrial protocol connectivity requires a gateway layer (e.g., Node-RED at the edge publishing to MQTT, which n8n then consumes)
• Self-hosting requires server administration skills; n8n explicitly recommends against it for non-technical users
• Queue mode (needed for high-availability) requires Redis and multiple instances — adds operational complexity
• Business plan (€667/mo) required for Git-based version control and multiple environments — community edition lacks these CI/CD features
• Execution-based pricing can become unpredictable for high-frequency polling workflows (e.g., polling every minute generates ~44,000 executions/month)