What is n8n?

What is n8n?

n8n is an open-source workflow automation platform — a tool for connecting apps and services into automated pipelines without writing glue code. You build workflows as visual node graphs: a trigger starts the run, nodes fetch, transform, branch, and act, and the result is an automation that runs the same way every time. Its official site is n8n.io and its documentation lives at docs.n8n.io.

If you've ever wished two of your tools would talk to each other — new form submissions should create a row somewhere, a webhook should kick off a multi-step process, a scheduled job should compile and send a report — n8n is the kind of tool built for exactly that.

The core idea

n8n's model is simple to grasp:

• Workflows as graphs. A workflow is a sequence of nodes: trigger → transform → API call → branch → action. You assemble it visually, and the execution path is explicit and auditable.
• Event-driven by default. Workflows start from triggers: webhooks (a URL that fires the workflow), schedules (cron-style), or events from connected apps.
• A large integration library. n8n ships connectors for a wide range of common services, plus generic building blocks — HTTP requests, webhooks, code nodes, database nodes — for everything not pre-built. The exact catalog changes over time; check the docs for the integrations you actually need.
• Deterministic execution. Given the same input, a workflow produces the same output. That property — the opposite of an AI agent's flexibility — is what makes n8n right for tasks where correctness and reproducibility matter more than improvisation. (See our article on workflow automation vs agents for the full comparison.)

It also has an AI-adjacent side: n8n workflows can call language models as nodes, which makes it a practical place to build automations that use AI where judgment is needed inside an otherwise deterministic pipeline.

The license: what "fair-code" means

n8n is open source, but under a specific model worth understanding. It's distributed under the fair-code "Sustainable Use License" — the fair-code definition is at faircode.io and n8n's legal pages at n8n.io/legal state the current terms.

What this means in practice:

• You can read, use, modify, and in many cases run n8n freely for personal and internal use.
• The license restricts commercial use beyond the stated terms — notably, offering n8n itself as a commercial service has conditions. This is the sustainable-use trade-off: the code is open, but the business model is protected.
• For a hosting provider, the consequence is direct: hosting n8n as a commercial service requires an appropriate commercial license arrangement with n8n. That's not optional or ambiguous, and it's a compliance question any would-be n8n host has to answer.

That last point is why HostAgentics treats it as a hard gate: n8n runtimes are only provisioned once the commercial license arrangement is confirmed (N8N_COMMERCIAL_LICENSE_CONFIRMED). If you're evaluating a hosted n8n provider, it's a fair question to ask them how they handle the license. If you're evaluating self-hosting n8n for your own organization, the fair-code terms still apply to commercial use — read them before you build a business process on it, and treat this article as orientation, not legal advice.

What n8n is good at

• Integration plumbing. Connecting systems that don't talk to each other: CRMs, spreadsheets, databases, email, messaging, webhooks.
• Scheduled and event-driven work. Anything that should happen on a schedule or in response to an event: nightly syncs, webhook processing, alert routing, report generation.
• Deterministic business processes. Ticket routing, data pipelines, approval flows — tasks where a fixed, auditable path is the point.
• Automation with an AI step. Using model nodes for classification, extraction, or drafting inside an otherwise fixed workflow.

What n8n is not

• Not a replacement for agents on open-ended tasks. If the task's steps can't be enumerated in advance — "read this mess and figure out what matters" — a fixed graph is the wrong tool; that's agent territory.
• Not a zero-maintenance product. Workflows break when upstream APIs change; integrations drift; version updates need testing. The _platform_ can be managed for you, but your workflows are yours to maintain.
• Not infinite. Hosted plans have hard limits on execution concurrency and resources — enforced caps, not billing meters, on HostAgentics.
• Not a single-vendor stack. n8n is one component in a stack that often includes agents, databases, and humans. It's excellent at being a component.

Running n8n

Your options mirror the general pattern: self-host on your own machine or server (full control, you own updates/backups/monitoring), or use a managed n8n — n8n's own cloud offering exists, and so do independent hosts. Managed hosting for n8n means someone else provisions the runtime, keeps it updated from tested versions, backs it up, monitors it, and restarts it — which is what HostAgentics provides: an isolated n8n runtime on its own branded domain with automatic HTTPS, daily provider snapshots with 7-day retention (14 with the Resource Boost add-on), automatic health monitoring and restart, and fixed EUR pricing with no overages.

Start at docs.n8n.io, build one workflow that replaces a task you currently do by hand, and let it run for a month. The editor is approachable, the docs are good, and the proof is in the automations you stop doing yourself.

Frequently asked questions

What is n8n used for?

n8n connects apps and services into automated workflows: a webhook fires a multi-step process, a schedule compiles and sends a report, a new form submission creates a row in your database. It's strongest for deterministic, auditable automations — anything where the steps are known in advance.

Is n8n free?

n8n is open source under a fair-code license (the Sustainable Use License). Self-hosting for personal and internal use is generally permitted; offering n8n itself as a commercial service has conditions, and a hosting provider needs a commercial license arrangement. Check n8n's legal pages for the current terms.

Can n8n run AI workflows?

Yes. n8n can call language models as workflow nodes, so you can add classification, extraction, or drafting steps inside an otherwise deterministic pipeline. For tasks that need a persistent, self-directing agent instead, see n8n vs AI agents.

What's the difference between n8n and an AI agent?

n8n runs a fixed, deterministic graph; an AI agent decides its own steps for open-ended tasks. They're complementary: use n8n for the parts that must not fail, and an agent for the parts that need understanding.

How much does it cost to host n8n?

It depends on whether you self-host (your server plus your time) or use managed hosting. HostAgentics n8n hosting runs an isolated n8n runtime at a fixed monthly price with automatic HTTPS, daily backups, monitoring, and updates — no infrastructure overages. See pricing for the current plans.