n8n

n8n 2.0 (Jan 2026) transforms the platform into a powerful orchestration layer for AI Agents, moving beyond linear automation. Native LangChain support enables Chains, Agents, Memory, and Vector Stores as first-class nodes. Any n8n workflow can be designated as a "Tool" that AI Agents call when needed—enabling modular, reusable agent skills. Persistent Memory (Redis, Postgres, in-memory) maintains context across executions. AI Workflow Builder generates workflows from natural language descriptions. Self-hosting the Community Edition is free with unlimited workflows, steps, and executions—your server's CPU is the only limit. Full data sovereignty for HIPAA, GDPR, SOC 2 compliance. Deploy via Docker, npm, or Kubernetes. 400+ native integrations with OpenAI (GPT-4o), Anthropic Claude, Google Gemini, HuggingFace, and local LLMs. Custom JavaScript/Python in Function nodes. Advanced node grouping, collapsible sections, improved execution tracing, and autosave (new in 2.0). Used by NASA, Adobe, and Twilio. Fair-code licensed—free to use with restrictions on reselling. Self-Hosted Community: Free (unlimited everything, infrastructure costs only ~$5-10/mo small, $200-500/mo production). Cloud Starter ($24/mo or $20/mo annual): 2,500 executions, 20 AI Builder credits. Cloud Pro ($60/mo or $50/mo annual): 10,000 executions, webhooks, priority support. Cloud Business ($667/mo annual): 40,000 executions, Git version control, SSO/LDAP, separate environments. Enterprise: custom pricing, dedicated support.