n8n: AI workflows you can actually see and control
AI 6 min read

n8n: AI workflows you can actually see and control

A fair-code automation tool that lets you drop into code anywhere and keep your data on your own servers.

Published Jun 30, 2026 Updated Jul 1, 2026

There is a particular kind of 9pm chore that quietly wears people down. A lead fills out a form on your site, and now someone has to copy their name into the CRM, drop a note in the sales channel, and set a reminder to follow up on Tuesday. None of it is hard. It is just tedious, repetitive, and very easy to forget at the end of a long day. n8n exists to make that whole chain happen on its own, and, more interestingly, to let you watch every step while it runs.

Quick facts

Pricing (as of 30 Jun 2026): free if you self-host the Community Edition. Cloud plans start at 20€/month (Starter, billed annually), with Pro at 50€/month and Business at 667€/month. Enterprise is custom. Every paid plan includes unlimited users and unlimited workflows.

Platforms: run it yourself with Docker on your own server, or use n8n's hosted cloud.

Open source? Sort of. The code is public on GitHub (around 194,000 stars), but it ships under a fair-code license rather than a classic open-source one. That distinction matters, and we will get to it.

What it actually is

n8n (say "n-eight-n", a squashed version of "nodemation") is a workflow automation tool. A workflow is just a sequence of steps you want to run automatically: when this happens, do that, then this. You build one on a visual canvas by dragging boxes, which n8n calls nodes, and drawing lines between them to show the order things should run in. Each node does one small job: grab an email, call an API (the way two pieces of software talk to each other), run a bit of logic, send a message to Slack.

The twist that has made n8n popular lately is how far it goes with AI. You can drop a language model into the middle of a workflow, build an agent (a setup where the AI decides which step to take next instead of you wiring every branch by hand), and pause for a human to approve an action before it actually fires. The selling point n8n keeps coming back to is visibility: you can see the agent's reasoning laid out on the canvas instead of trusting a black box.

The problem it solves

Picture a small company that just closed a new sale. Normally a person now does four little jobs by hand: pull the customer's details out of the CRM, write them a welcome email, create their setup task in the project tracker, and post a heads-up in the team channel. Four chores, every single time, all copy-and-paste.

With n8n, one trigger handles the lot. The new deal in your CRM kicks off the workflow. n8n fetches the details, has an AI model draft a warm, personal welcome note, opens the task in your project tool, and posts the note for the team. The salesperson closes the deal and moves on. The busywork runs itself in the background, and if anything breaks, the logs show you exactly which step tripped.

Who it is for

n8n is aimed squarely at technical teams. Developers love it because they are never trapped in the visual editor: when a drag-and-drop step cannot do what they need, they can write a few lines of JavaScript or Python right inside the workflow and carry on. Operations folks, IT, and security teams use it to automate the internal grind, like onboarding a new hire's accounts or enriching a pile of alert tickets.

If you are a complete non-coder who wants the simplest possible point-and-click experience, n8n can feel busy. The canvas rewards people who are comfortable thinking in terms of data moving from one step to the next. That is a fair trade for the power it hands back, but it is worth knowing going in.

Why it stands out

A few things make n8n feel different from the usual automation tool.

You can drop into code anywhere. Most visual builders box you in the moment your idea gets slightly unusual. n8n lets you switch to real code in any step and switch back, so a tricky bit of logic does not force you to abandon the whole approach.

You can run it on your own machines. Self-host with Docker and your workflow data stays on your servers, which matters a lot when that data is customer records or security alerts.

Billing is per execution, not per step. One full run of a workflow counts as one execution, no matter whether it has three steps or thirty. In tools that charge for every individual action, a single busy workflow can quietly rack up thousands of billable operations. n8n's model is easier to predict.

Here is the honest catch, and it is a real one. n8n calls itself fair-code, not open source, on purpose. You are free to use, modify, and self-host it for your own internal business or personal use, at no cost. What you cannot do is repackage it, bake it into a product you sell, or run it as a paid service for outside users without a commercial license. So it is genuinely free right up until the day you try to build a business on top of it. For most teams automating their own internal work, that line never gets crossed. If your plan is to resell automation to clients, read the license first.

How to try it

The fastest route is the cloud free trial, which needs no credit card. You sign up, pick a template, and watch something run within a few minutes.

If you would rather keep everything on your own server, the self-hosted version takes two commands with Docker:

docker volume create n8n_data
docker run -it --rm --name n8n -p 5678:5678 -v n8n_data:/home/node/.n8n docker.n8n.io/n8nio/n8n

Then open localhost:5678 in your browser. Either way, do yourself a favor and start from one of the 10,000-plus prebuilt templates instead of a blank canvas. Seeing a finished workflow and tweaking it teaches you the tool far faster than staring at an empty page.

Takeaway

n8n is the rare automation tool that does not make you choose between a friendly visual builder and the freedom to write real code. For a developer or a technical team that wants to wire AI into their actual work, keep the data on their own servers, and still understand what every step is doing, it is one of the most capable options going. Just go in clear-eyed about the two catches: a canvas that expects a little technical comfort, and a license that is free to use but not free to sell.

Share this article
Weekly digest

One email a week.
The five things that mattered.

Friday mornings. No hype. Unsubscribe anytime.

By subscribing, you agree to our Terms of Use and Privacy Policy.