n8n workflow automation — connected nodes with arrows
FIELD GUIDE · AUTOMATION

What Is n8n? Build AI-Powered Automation Workflows (2026 Guide)

2026-04-03·9 min read·Automation

n8n is a source-available workflow automation platform that lets you connect APIs, run logic and orchestrate AI agents — either as a hosted service or self-hosted on your own infrastructure. This guide covers what n8n is, how it works, where it fits compared with Zapier and Make, when to self-host, and the practical patterns for using it as an AI agent runtime.

What is n8n?

n8n (pronounced "n-eight-n", short for "nodemation") is a workflow automation platform that lets you build, schedule and run workflows by connecting nodes — visual blocks that represent triggers, actions or logic — across hundreds of APIs and services. It is source-available under the Sustainable Use License, meaning you can self-host for free for internal use, and a managed cloud offering plus enterprise edition are also available.

How does n8n work?

A workflow in n8n is a graph of nodes connected by lines. A trigger node fires (webhook, schedule, file change, new email, MCP call), data flows through transformation and action nodes, and the workflow ends when there's nothing left to do. Each node receives input as JSON, performs its work and emits output as JSON. Code nodes let you drop into JavaScript or Python when the built-in nodes aren't enough.

What can you build with n8n?

  • AI agents — chain LLMs with tools, memory and retrieval; multi-agent workflows.
  • Internal automation — sync data between SaaS tools, enrich CRM records, file tickets automatically.
  • ETL and reporting — pull from APIs, transform with code nodes, push to data warehouse.
  • DevOps and SecOps — alert routing, on-call workflows, automated remediation steps.
  • Customer-facing flows — webhook handlers, AI chat backends, document processing.
  • Compliance — evidence collection on schedule, audit log enrichment, control validation.

n8n and AI agents

n8n ships with a first-class set of LangChain-based nodes: chat models, embeddings, vector stores, document loaders, memory, tools and an Agent node that executes a tool-using loop with the model of your choice (OpenAI, Anthropic, Google, Mistral, Ollama for local models). The visual canvas makes it easy to wire an agent to your real data and real APIs without writing the orchestration code yourself, while keeping the option to drop into code when needed.

n8n vs Zapier vs Make

  • Pricing — n8n charges per workflow execution; Zapier per task; Make per operation. At scale, n8n is typically 50–80% cheaper.
  • Data control — n8n can be self-hosted; Zapier and Make are cloud-only.
  • Complex logic — n8n's branching, looping, error handling and code nodes are significantly more powerful.
  • Integration breadth — Zapier has the largest catalog (~7,000+); n8n covers 500+ but includes HTTP and code as fallbacks.
  • AI features — all three offer AI nodes; n8n's first-class agent support is the most flexible.
  • Ease of use — Zapier is the friendliest for non-technical users; n8n assumes more familiarity with APIs and JSON.

Self-hosting n8n

Self-hosting is the main reason technical teams choose n8n. Run it with Docker, Docker Compose or Kubernetes; persist workflows in PostgreSQL; put it behind your VPN or reverse proxy; use your own SSO. The Sustainable Use License allows free self-hosted use for internal business purposes — embed-in-a-product use requires a commercial license.

Security considerations for n8n

  • Treat n8n as a privileged automation hub — it holds credentials for everything it integrates with.
  • Use the encrypted credentials store, never plain-text in workflow JSON.
  • Restrict access via SSO, MFA and RBAC (enterprise edition).
  • Log workflow executions and back up regularly.
  • Audit any community nodes before installing — they run arbitrary code.
  • Network-isolate the n8n host from production data planes when possible.

When to choose n8n (and when not to)

Choose n8n when: you have a technical team, you need self-hosting for compliance or cost, you build complex branching workflows, or you want a flexible AI agent runtime. Don't choose n8n when: your users are non-technical line-of-business operators, your needs are 100% covered by Zapier's catalog, or you can't dedicate someone to maintain a self-hosted service.

Is n8n really free?+

Self-hosted n8n is free for internal business use under the Sustainable Use License. Embedding in a product you sell, or providing n8n-as-a-service to others, requires a commercial license.

Can n8n run AI agents?+

Yes. n8n ships with native LangChain-based nodes for chat models, embeddings, vector stores, memory and an Agent node — across OpenAI, Anthropic, Google, Mistral and local models via Ollama.

How does n8n compare to LangGraph or CrewAI for agents?+

n8n is visual-first and integration-rich — best when the agent needs to call many real-world tools. LangGraph and CrewAI are code-first frameworks — better when you need granular control over the agent loop itself.

What does n8n cost in production?+

Self-hosted: only your infrastructure (€20–€100/month for typical workloads). Cloud Starter from €20/month; Pro €50/month; Enterprise custom. Compared to Zapier at scale, savings of 50–80% are common.

Is n8n secure enough for sensitive data?+

Self-hosted n8n with SSO, encrypted credentials, network isolation and audited community nodes is appropriate for sensitive data. Treat it as a privileged automation hub and harden accordingly.

NEXT STEP

CyberAce designs, builds and operates secure n8n deployments — including AI agent workflows for SecOps and compliance.

Talk to CyberAce