n8n

n8n is a workflow automation and AI agent platform built for technical teams who need the flexibility of code without sacrificing the speed of a visual builder. It operates under a fair-code license — source code is always visible and self-hostable, while commercial cloud and enterprise offerings layer on managed infrastructure and governance features. The platform is actively developed, with the GitHub repository showing 188,754 stars and the latest release at version 2.21.4 as of May 2026.

What It Is

n8n (pronounced "n-eight-n", short for "nodemation") is a node-based workflow automation tool that lets users connect apps, transform data, and build AI agent pipelines through a canvas-style editor. Unlike purely no-code tools, n8n allows JavaScript or Python code at any step, npm package imports on self-hosted instances, and custom API calls via HTTP request nodes. The result is a platform that sits between traditional iPaaS tools and full custom development — visual enough for rapid prototyping, deep enough for production-grade automation.

How the Visual-Plus-Code Model Works

The core building block is the "node" — a configurable step that represents a trigger, an app action, a data transformation, or an AI operation. Users wire nodes together on a canvas to form workflows. Key mechanics include:

• Diverse triggers: webhooks, schedules (cron), app events, chat inputs, and event streams
• Data routing: switch/if nodes, loops, merge nodes, and deduplication
• Code nodes: write arbitrary JavaScript or Python inline; expressions support a JS-based templating language called Tournament
• Fast iteration: re-run single steps, pin previous execution data, replay mocked inputs without waiting for external systems
• AI nodes: built-in LangChain integration for multi-step agents, RAG pipelines, and model-agnostic LLM calls

Every node shows its inputs and outputs side-by-side with settings, reducing context switching during debugging.

AI Agent and RAG Capabilities

n8n positions AI workflow building as a first-class use case. The platform supports multi-agent setups where sub-agents handle specialized tasks, and RAG (retrieval-augmented generation) pipelines that connect vector stores and document sources to LLMs. Users can connect any cloud or locally-hosted model, inspect every reasoning step on the canvas, and enforce human-in-the-loop approval gates before AI actions execute. MCP (Model Context Protocol) support is listed as available for connecting to external tool registries. The enterprise page highlights AI governance features including evaluations, guardrails, and structured input/output enforcement.

Deployment Model and Licensing

n8n offers two primary deployment paths:

• Self-hosted: Deploy via Docker (docker run) or Kubernetes in minutes; run air-gapped on a private network; source code available on GitHub under the Sustainable Use License and n8n Enterprise License (fair-code, not OSI-approved open source)
• Cloud (n8n.cloud): Managed hosting with data stored in Frankfurt, Germany; SOC 2 audited

The community edition is free to self-host. Business and Enterprise plans require a license key that pings n8n's license server daily to track production execution counts. The GitHub README notes that enterprise license inquiries go to license@n8n.io.

Enterprise Governance and Scale

The enterprise page states that n8n claims 34% of Fortune 500 companies use its advanced security and DevOps features. Enterprise-specific capabilities include:

• SSO, SAML, LDAP, and user provisioning via identity providers
• Role-based access control (RBAC) with project-level isolation
• External secret store integration (HashiCorp Vault, AWS Secrets Manager, Azure Key Vault, GCP)
• Git-based source control with workflow diffs between environments
• Log streaming to SIEM/observability stacks (e.g., Datadog)
• Queue mode and multi-main instance setups for horizontal scaling; the features page states n8n handles up to 220 workflow executions per second on a single instance
• Audit logs, workflow history, and an Insights dashboard showing execution counts, failure rates, and time-saved metrics

Update: Version 2.21.4

The latest GitHub release is n8n@2.21.4, published on May 19, 2026. The repository was last pushed to on May 20, 2026, indicating active weekly release cadence. The project was created in June 2019 and has accumulated 188,754 stars and 57,832 forks, making it one of the most-starred workflow automation repositories on GitHub. The pricing page lists an AI Workflow Builder feature (with per-plan credit allocations) and notes that AI Workflow Builder is "coming soon" for the Business self-hosted tier, signaling ongoing AI feature expansion.