n8n
Visual workflow and AI agent platform you can self-host and inspect step by step
- A visual workflow builder that accepts code steps when a task needs more control.
- AI agents that run inside a workflow, with live views of current runs and records of past ones.
- Guardrails for agent behavior, including structured inputs and outputs, human approval steps and rule-based branches.
- More than 500 native integrations, plus connections to APIs and other agents through MCP.
Best for: Developers and technical teams that automate internal processes. · IT, security and DevOps teams that need self-hosting and governance. · Operations teams adding AI agents to existing workflows.
Introduction
n8n is a workflow automation platform for connecting apps, APIs and AI agents in one place. Workflows are built on a visual canvas, and any step can be opened to see its input, output and configuration. That visibility is the core idea: agents and automations should be inspectable by the people who depend on them. Teams can run n8n in the cloud or host it themselves.
Key Features
- A visual workflow builder that accepts code steps when a task needs more control.
- AI agents that run inside a workflow, with live views of current runs and records of past ones.
- Guardrails for agent behavior, including structured inputs and outputs, human approval steps and rule-based branches.
- More than 500 native integrations, plus connections to APIs and other agents through MCP.
- Self-hosting on your own infrastructure or with Docker.
- Version history, audit logs, SSO with SAML, LDAP, role-based permissions and Git-based control.
- A one-step recovery option that re-runs only the failed part of a workflow.
Who is it for
- Developers and technical teams that automate internal processes.
- IT, security and DevOps teams that need self-hosting and governance.
- Operations teams adding AI agents to existing workflows.
- Enterprises that require SSO, audit logs and role-based access.
What can you do with n8n
- Automate multi-step processes across business applications.
- Add an AI agent that reasons over data and calls the tools it needs.
- Put an approval step in front of actions that carry risk.
- Replay a past run with mock data to test a change before release.
- Self-host the platform to keep workflow data on your own infrastructure.
- Give other agents and coding assistants access through MCP.
How does n8n work
- Start from a blank canvas or a template and add a trigger.
- Connect applications, APIs and data sources as nodes on the canvas.
- Add agent nodes, choose a model and define the tools and guardrails the agent may use.
- Insert approval or branching steps where a person or a rule must decide.
- Run the workflow, open any step to inspect its data, then fix and re-run the failed part.
Pricing
- The Community Edition can be self-hosted at no listed cost.
- Starter is 20 euros per month billed annually, with 2.5K executions and unlimited steps.
- Pro is 50 euros per month billed annually, with 10K executions.
- Business is 667 euros per month billed annually, with 40K executions, and is self-hosted only.
- Enterprise is custom priced, with custom executions and either hosted or self-hosted deployment.
FAQ
Is n8n open source?
n8n publishes a Community Edition that can be self-hosted at no listed cost. The pricing page lists it alongside the hosted Starter, Pro, Business and Enterprise plans.
Can I host n8n myself?
Yes. Self-hosting is supported on your own infrastructure or with Docker, and the Business plan is self-hosted only. Hosted cloud plans are also available.
Do I need to write code to use n8n?
No. Most workflows can be assembled visually. Code steps are available when a task needs custom logic, and developers can use Git-based control for larger deployments.
What counts as an execution?
An execution is one run of a workflow. Paid plans set a monthly execution allowance while leaving the number of steps inside each workflow unlimited.
Can n8n run AI agents?
Yes. Agents can run as part of a workflow, call tools, and be wrapped in approvals or rule-based checks. Teams can watch a run live or review its history afterward.
