Skip to main content

n8n is a workflow automation platform that lets you visually connect apps and services, then drop into code whenever the visual builder isn't enough.

It is for anyone who repeats the same multi-step task across different apps. It especially suits people who expect to need more flexibility than a fully no-code tool gives them.

Without it, someone has to manually repeat the same steps across different apps and services every time a task needs doing.

Nodes and credentials

A node is one step in a workflow. Each node performs a single action, such as fetching data, sending data, or transforming it, against an app, a service, or your own code. Many nodes need credentials, private login details like an API key, to connect to the app or service they talk to.

A NASA node fetches recent solar flare data from NASA's public API. To do that, it needs a NASA API key stored as a credential, so n8n can authenticate the request on your behalf.

The technical details

n8n's own docs describe credentials as "private pieces of information issued by apps and services to authenticate you as a user."

A credential lets n8n "connect and share information between the app or service and the n8n node."

What a specific credential requires depends on the app or service involved. It might be an API key, an OAuth login, or a token.

Nodes also expose settings like Always Output Data and Execute Once, which control what happens to the data passing through them.

Workflows

A workflow is a set of nodes connected together on a canvas, built to carry out a task from start to finish. You can run a workflow by hand or have it start automatically.

A workflow checks a NASA API on a schedule and tests the result against a condition. It then sends one of two different reports, depending on what it finds.

The technical details

n8n's tutorial for building a first workflow shows the pattern directly. Add a trigger node, add action nodes after it, add logic with an If node to branch the workflow, then send the result somewhere. Once a workflow works, you publish it, so it keeps running automatically on its trigger rather than only when you press Execute Workflow by hand.

Triggers vs. actions

Every node is either a trigger or an action. A trigger starts a workflow when something happens. An action is a task the workflow performs once it's running.

A Schedule Trigger starts a workflow every Monday at 9am. After it fires, an action node fetches data, and another action node sends it somewhere.

The technical details

n8n's own docs define both terms directly. "Triggers start a workflow in response to specific events or conditions in your services."

When you select one, n8n adds a trigger node with that operation pre-selected. "Actions are operations that represent specific tasks within a workflow."

Actions manipulate data, act on external systems, or trigger events elsewhere. Triggers can fire on a schedule or a webhook call. They can also fire on an event in a connected app, such as a new email.

Executions

An execution is one complete run of a workflow, from its trigger to its last node. It doesn't matter how many nodes the workflow has, ten or two. A single run still counts as one execution, and this is the unit n8n uses to measure usage and bill for its paid plans.

A 20-node workflow triggers on a new email, extracts data, enriches it, updates a CRM, and posts a Slack message. It still counts as one execution, not 20.

The technical details

n8n contrasts this directly with Zapier and Make on its own comparison pages. Zapier "charges for every single task (or step) in your workflow," and Make bills "every click, filter, and API call" as a separate operation.

n8n's pages state plainly that it "doesn't count tasks. It only charges for workflow executions." Execution counting has some documented edge cases.

A sub-workflow called from a main workflow counts as its own separate execution. Manual test runs in the editor count the same as production runs.

A polling trigger consumes one execution per check cycle even if nothing new is found. Unused executions don't roll over to the next month.

Problems it solves

Doing the same multi-step task across apps, by hand, every time

Repetitive work across different apps normally means a person doing the same steps over and over. An example is copying form submissions into a spreadsheet and then emailing a confirmation. A workflow does it automatically once, and then keeps doing it on its own trigger.

No-code tools that hit a wall as soon as logic gets complex

Fully visual, no-code automation tools work well for simple, linear tasks. n8n's own comparison pages argue that this breaks down once a workflow needs branching, looping, or custom logic. n8n lets you write JavaScript or Python directly inside a node when the visual builder alone isn't enough, without leaving the workflow or switching tools.

Paying per step instead of per result

n8n says Zapier and Make use task-based and operation-based pricing. That model charges more as a workflow gets more complex, since every additional step is a separate billable unit.

n8n instead charges per execution, one full run of the workflow regardless of how many nodes it contains. Adding steps and logic doesn't multiply the bill.

Wanting to self-host and keep control of your own automation data

A fully managed, cloud-only automation tool means your workflow data lives on someone else's infrastructure by default. n8n can be self-hosted for free instead, using the Community edition on your own servers or private cloud.

It comes with unlimited executions and almost the complete feature set. A team that needs to keep data in-house has that option built in rather than bolted on.

Competitors

Zapier, Make

Pricing

Free self-hosted Community edition, plus paid Cloud tiers billed by monthly executions. Starter plan starts at €20 a month, billed annually, for 2,500 workflow executions

Business model

Private company. $180M Series C, October 2025, led by Accel, $2.5B valuation

Links

License
Sustainable Use License (source-available, fair-code); separate proprietary n8n Enterprise License for paid features

Going deeper

What happens when a workflow runs?

n8n's own tutorial for building a first workflow walks through the real mechanism step by step.

Trigger, then actions, in sequence

A workflow starts with a trigger node, added first. n8n's tutorial uses a Schedule Trigger set to run weekly. A trigger can just as easily fire from an incoming webhook or an event in a connected app instead.

Once the trigger fires, action nodes run one after another. Each one performs a task, such as fetching data from an API, checking a condition, or sending data somewhere else.

The tutorial's example fetches solar flare data from NASA's public API. It then uses an If node to branch the workflow into two paths, based on the classification of the data returned. It sends a different report down each path.

Expressions carry data between nodes

Static values aren't always enough. n8n's expressions let you insert dynamic data into any field instead of typing a fixed value.

The tutorial's own example generates a date seven days before the current date. That way, the NASA node fetches only the past week's flares rather than its default 30-day window.

A later step in the same tutorial drags a field named classType straight from one node's output into another node's input. n8n turns this into the reference {{$json["classType"]}} automatically.

What happens when a node fails

Node settings control what happens on error. Retry On Fail reruns a failed node until it succeeds.

On Error offers three modes. Stop Workflow halts everything, and Continue moves on using the last valid data.

Continue (using error output) passes the error itself to the next node, so the workflow can handle it explicitly rather than just carrying on blind.

What happens when the visual builder isn't enough?

n8n's own homepage frames this directly as a choice it doesn't force you to make. "Code when you need it, UI when you don't."

Where a built-in node can't do what you need, you can write JavaScript or Python directly inside a Code node. It sits in the same workflow, next to the visual nodes around it.

n8n's GitHub README describes this as combining visual building with custom code and npm packages, treating code as part of the tool rather than a separate escape hatch outside it. Wikipedia's independent account confirms the same pattern: where visual nodes fall short, users write JavaScript or Python in a code node instead of switching to an entirely different system.

Self-hosting vs. n8n Cloud

n8n's docs draw a clear line between an edition, the software variant you run, and a plan, the subscription tier you pay for.

The free, self-hosted Community edition

Every self-hosted installation runs on the same underlying product. Without a license key, that's the Community edition, free indefinitely, with what n8n's own docs call "almost the complete feature set."

Registering it with an email unlocks a few extra free features, such as folders for organizing workflows, still at no cost. n8n can be self-hosted using npm, Docker, or a cloud provider like AWS or Azure.

Paid plans, cloud and self-hosted

n8n Cloud is the fully managed option. You sign up and build, and n8n handles servers, updates, and security patches.

n8n's own pricing page puts the cloud Starter plan at €20 a month, billed annually, for 2,500 executions a month with unlimited steps.

Pro runs to €50 a month, billed annually, for 10,000 executions. Business runs to €667 a month, billed annually, for 40,000 executions, and is self-hosted rather than running on n8n's own infrastructure, adding features like SSO and Git-based version control.

Enterprise pricing is custom, with a custom number of executions, available either hosted by n8n or self-hosted. Companies under 20 employees can also apply for a Start-up Plan, which n8n's pricing page states gets 50% off Business.

Why a team might pick one over the other

n8n's own Community edition and Business/Enterprise licensed editions cover self-hosted deployment for teams that need to keep infrastructure and data under their own control. Cloud removes that operational burden entirely.

n8n's docs put the choice plainly. Self-host for free with almost the full feature set, or use Cloud if you'd rather not manage infrastructure at all.

What does the fair-code license allow?

n8n's source code is visible and free to self-host. The Sustainable Use License draws a specific line around what you can build on top of it.

What the license permits and restricts

n8n created the Sustainable Use License itself in 2022. It modeled the license on the Elastic License 2.0, replacing an earlier Apache 2.0 plus Commons Clause arrangement.

It grants a free right to use, modify, and redistribute the software, with three limits. You can only use or modify it for your own internal business, non-commercial, or personal purposes. You can only redistribute it free of charge for non-commercial purposes.

And you can't remove n8n's own licensing or copyright notices.

In practice, n8n's own FAQ says this means all use is allowed unless the value of what you're selling comes substantially from n8n itself. White-labeling n8n and reselling it to your customers, or hosting it and charging people to access it, aren't allowed.

Using n8n internally to sync your own company's data and building an integration node for your own product are both allowed. So is offering paid consulting to build n8n workflows for clients.

Why n8n calls this fair-code, not open source

Because the license restricts use, n8n doesn't call itself open source under the Open Source Initiative's definition, which doesn't allow use restrictions at all. n8n coined the term "fair-code" instead, to describe a model that keeps source code open and self-hostable while carrying commercial restrictions on top.

Enterprise customers who need something the Sustainable Use License doesn't cover get a separate proprietary agreement instead, the n8n Enterprise License.