n8n is a workflow automation and AI agent platform from Berlin that charges per full workflow execution, however many steps it has. The self-hosted Community Edition is free under n8n's Sustainable Use License. n8n Cloud costs €20 a month billed yearly for 2,500 executions on Starter, €50 for 10,000 on Pro and €667 for 40,000 on the self-hosted Business plan, with unlimited users and workflows on every plan. When a Cloud plan hits its execution limit, new executions fail immediately rather than queuing until the next month or an upgrade.
Reviewed by EdGrows·Updated Oct 1, 2026This analysis is based on documentation, public user reports, and vendor materials — not yet on our own hands-on testing. How we rate
n8n started in 2019 as a visual workflow tool for developers and grew into one of the most popular automation platforms, with more than 200,000 stars on GitHub. Its homepage now leads with agents: "AI agents and workflows you can see and control." You build on a canvas of nodes, drop into JavaScript or Python where a node isn't enough, and can trace every step an agent took.
Two things set it apart from Zapier and Make. It charges per execution, a full run of a workflow, rather than per step, so long workflows don't multiply the bill. And you can run the whole thing on your own infrastructure, which matters for teams with data that cannot leave their network.
The self-hosted Community Edition is free, with unlimited workflows, users and executions. You run it on your own server, often with Docker, and pay only for hosting, which for a small team can be a few euros a month on a basic cloud server. Running locally also lets n8n reach files and internal systems that hosted automation tools cannot.
The licence is the detail many guides get wrong. n8n's code is public, but it is released under n8n's Sustainable Use License, which n8n describes as fair-code rather than open source. Internal business use is free; offering n8n as a hosted service to others, or building a competing product on it, needs a commercial agreement.
n8n's pricing page on annual billing on October 1, 2026. n8n shows its prices in euros.
| Plan | Billed yearly | Executions a month | Hosting | What it adds |
|---|---|---|---|---|
| Community Edition | Free | Unlimited | Your server | The core platform, all integrations |
| Starter | €20 a month | 2,500 | n8n Cloud | Agents, MCP and AI nodes, 5 concurrent executions, 1,600 AI Assistant credits |
| Pro | €50 a month | 10,000 | n8n Cloud | Up to 50 concurrent executions, admin roles, evaluations, workflow history, up to 9,600 credits |
| Business | €667 a month | 40,000 | Self-hosted | SSO with SAML and LDAP, multiple environments, Git version control, queue mode |
| Enterprise | By quote | Custom | Cloud or self-hosted | 200+ concurrent executions, external secrets, log streaming, SLA support |
Every plan includes unlimited users, unlimited workflows and every integration. Monthly billing costs about 17% more, and companies with fewer than 20 employees can apply for a start-up plan with 50% off Business. Starter workflows can run for up to 5 minutes, Pro and Enterprise for up to 40.
This is the part to plan for. n8n's help center says that when a Cloud plan reaches its monthly execution limit, executions are not queued: each new run fails immediately with the error "Execution limit reached. Consider upgrading your plan." Workflows stay active and triggers keep firing, so every webhook or schedule after the cap produces a failed run, and those runs are not retried later.
Normal service resumes when the limit resets at the start of the calendar month or when you upgrade. n8n offers a one-time courtesy reset for Starter and the 10,000-execution Pro tier, and monthly subscribers on the larger Pro tier can add executions themselves, up to 500,000 a month. For business-critical automations, size the plan with headroom or self-host.
n8n's AI features are aimed at people who want to see what an agent does. The AI Agent node connects a model to tools and memory inside a workflow, every step of its reasoning is traceable on the canvas, and human approval can be required before an agent calls a tool. Paid Cloud plans include AI models that work without your own API keys, plus AI Assistant credits for help building workflows.
MCP support runs both ways: the MCP Server Trigger turns your workflows into tools that assistants like Claude can call, and the MCP Client node lets n8n agents use tools from other servers. Evaluations test agent output against expected results, and self-hosted users can trace runs in LangSmith or, in preview, OpenTelemetry.
Zapier connects more apps with less setup and suits non-technical teams, but charges per task, so long or AI-heavy workflows cost much more. Make sits in between, with a visual builder, credits per module action and prices from $9 a month. n8n is the choice for technical teams that want code, self-hosting and predictable costs on complex workflows. Our comparisons of n8n vs Zapier and n8n vs Make go deeper.
The self-hosted Community Edition is free to run on your own server, with unlimited workflows and executions, under n8n's Sustainable Use License. You pay only for hosting. n8n Cloud, the hosted version, starts at €20 a month billed yearly for 2,500 executions and comes with a free trial that needs no credit card.
Not in the strict sense. n8n's source code is public on GitHub, where it has more than 200,000 stars, but it is licensed under n8n's own Sustainable Use License, which n8n calls fair-code. You can use and modify it for internal business purposes, while reselling it as a hosted service or competing product needs a commercial agreement.
Yes. The Community Edition runs on your own computer or server, typically with Docker or npm, and can reach local files and internal systems that cloud automation tools cannot. Teams that need SSO, Git-based version control or multiple environments on their own infrastructure use the paid self-hosted Business or Enterprise plans.
Billed yearly, Starter costs €20 a month for 2,500 executions and Pro €50 for 10,000, and monthly billing costs about 17% more. Business, at €667 a month for 40,000 executions, is a self-hosted plan with SSO, environments and Git. n8n shows prices in euros, and companies under 20 employees can apply for 50% off Business.
New executions fail immediately with an error asking you to upgrade; they are not queued, and failed runs do not re-run later. Workflows stay active, so they work again after the monthly reset or an upgrade. n8n offers a one-time courtesy reset on Starter and Pro, and some monthly Pro subscribers can add executions themselves.
Yes. n8n includes an MCP Server Trigger, which exposes your workflows as tools that AI assistants such as Claude can call, and an MCP Client node that lets n8n agents use tools from other MCP servers. Instance-level MCP and human approval for tool calls are also listed on its plans.
For technical teams building complex or AI-heavy workflows, often yes: n8n charges per full execution rather than per step, supports code and self-hosting, and costs far less at volume. Zapier connects more apps out of the box and is easier for non-technical users. Our n8n vs Zapier comparison works through the trade-offs.
Anthropic's assistant, built around long-context reasoning and code. Strong at $20; the usage limits are the unpublished part.
Google's assistant, bundled with storage and Workspace access. Plus is $4.99, Pro $19.99 with a 1 million token context, and Ultra from $99.99.
Visual automation with 3,000+ apps, billed in credits: a free plan with 1,000 a month, then $9 to $29 a month billed yearly for 10,000 credits.