n8n
Fair-code workflow automation with native AI — self-hosted or cloud
📋 Overview
n8n is the fair-code workflow automation platform with native AI capabilities — combining visual workflow building with custom JavaScript/Python, 400+ integrations, and self-hosting. Unlike Zapier (closed-source SaaS) or Apache Airflow (code-only), n8n offers both a visual node editor and full code extensibility, letting teams drag-and-drop simple automations while developers inject custom logic where needed. The AI agent features natively call any LLM or MCP server, with vector memory and RAG nodes built in. Its fair-code license (Functional Source License, built on Apache 2.0) lets you self-host for free with no usage caps; the cloud tier adds managed hosting, SSO, and support.
✨ Key Features
- •196K+ GitHub stars — largest automation project on GitHub
- •400+ native integrations — Zapier-class breadth
- •Fair-code license (Functional Source License 1.1) self-hostable for free
- •Native AI agent nodes: LLM calls, vector memory, RAG, tool calls
- •MCP server and MCP client built into the workflow engine
- •Custom Python and JavaScript nodes for unlimited extensibility
- •Cloud SSO, RBAC, audit log, and HIPAA compliance on higher tiers
🎯 The Problem It Solves
Enterprises need workflow automation that balances developer power with business-team accessibility — without hitting vendor lock-in or per-run pricing that cripples scale. n8n gives non-technical users visual flows, developers custom-code nodes, and ops teams full self-hosting control with AI agents built natively into the workflow engine.
🔧 How It Works
Workflows are built by chaining nodes — triggers (webhooks, schedules, email), actions (APIs, DB queries, AI calls), and logic (if/else, loops, merge). The native AI node connects to any LLM API (OpenAI, Anthropic, Gemini, etc.) or directly to an MCP server, with vector memory nodes for RAG. Data flows between nodes as JSON, inspectable at every step. Workflows can be version-controlled via git, and the fair-code license lets you build custom nodes for internal use.
🚀 Installation & Quick Start
Installation
docker run -it --rm --name n8n -p 5678:5678 n8nio/n8nQuick Start
- docker run ...
- Open localhost:5678
- Create a workflow
✅ Pros
- •Self-hosted is completely free with no execution caps
- •Visual drag-and-drop AND custom code — both paradigms work
- •Native AI nodes bypass third-party automation wrappers
- •MCP support is built-in — not a plugin
- •400+ integrations out of the box
- •Fair-code license prevents vendor lock-in while funding development
- •Active community: templates, tutorials, workshops
❌ Cons
- •Self-hosting requires ops investment (Redis, PostgreSQL, NodeJS)
- •Node editor becomes cluttered past 50+ nodes
- •Complex error handling requires JavaScript knowledge
- •Cloud execution limits can blow budgets on high-volume automations
- •Enterprise tier pricing opaque — requires sales conversation
💬 Practitioner Verdict
“n8n is the default workflow automation platform in 2026 — the fair-code model threads the needle between open-source flexibility and cloud-service convenience. The native AI and MCP support make it the backbone of most agentic automations I see in production. The trade-off: self-hosting requires ops investment, the node editor can feel slunky at 50+ nodes, and complex error handling requires JavaScript.”
Self-Hosted (Free)
Open source, MIT/Apache licensed. Run it yourself.
⭐ Star & Clone on GitHubFree forever. Your infrastructure, your data.
n8n Cloud Starter
Managed hosting with 5K executions/mo
- Managed hosting
- 5K executions
- Community support
n8n Cloud Pro
25K executions + SSO + insights
- 25K executions
- SSO
- Workflow insights
Enterprise
Unlimited, VPC, HIPAA, dedicated support
- Unlimited executions
- VPC/HIPAA
- Dedicated support
Deployment Options
Ways to run n8n in production
📊 Specifications
- Language
- TypeScript
- License
- Fair-code (Functional Source License 1.1)
- Platform
- Linux, macOS, Windows
- Supported Models
- REST API, CLI
💰 Pricing Reality
Self-hosted is 100% free with no usage caps (fair-code license). Cloud: Starter 0/mo (5K workflow executions), Pro 0/mo (25K executions, SSO, insights), Enterprise custom (unlimited, VPC, HIPAA). All LLM API usage billed separately to your provider.