What are the three ways to pay for n8n work?
A single workflow or a fix starts at $49, and a build around one real process starts at $499, each as one fixed price agreed in writing after a free call. Keeping it alive afterwards starts at $149 a month. Those are the same published starting points as every other kind of work on the services and pricing page.
The number is the easy part. What decides whether you got value is which of the three arrangements you bought, because each one puts the risk of a slow or messy job on a different person.
| Way to pay | What you are buying | Who carries the risk | What it hides |
|---|---|---|---|
| By the hour | Time, not an outcome | You: a slow job costs more | How long it should have taken, and whether the hours were spent on the problem or on the tooling |
| One fixed price | A written list of what done means | The developer: a slow job costs them | Anything not on the list, which is why the list has to be specific before you agree |
| A monthly amount | Someone watching it and fixing it | Shared, and renewable | Whether the month included new work or only repairs, and what happens to the workflows if you stop |
Which one fits your job?
Match the arrangement to how well the job can be described, not to how big it feels. A vague job bought by the hour is how a small automation becomes a four month invoice.
- You can write the job down: one fixed price. What goes in, what comes out, what happens when it fails, and what is explicitly not included.
- You cannot write it down yet: buy a small fixed piece of work that produces the plan. A short audit of the current process is a deliverable, and it is cheaper than an open hourly engagement.
- It is already running and must keep running: a monthly arrangement. This is the one most small businesses skip, and it is the one that decides whether the automation is still working in month nine.
- Something is broken right now and you do not know why: a fixed-price fix if the symptom is clear, otherwise a short paid diagnosis first. An hourly rate on an unknown fault is an open cheque.
- You want a person on call every week: that is a retainer, and it should say how many hours or how many changes it covers. Ask what happens in a month where you use none of it.
What does a year of n8n cost, rather than a month?
Budget the year, because the build is paid once and everything else is paid twelve times. The categories and the cloud versus self-hosted choice are already written up in how to hire an n8n developer, so this is about how the total moves rather than a second list of the same items.
Three things make an n8n bill grow after launch, and none of them is visible on the day you sign:
- Volume you did not plan for. A workflow that runs on every order is fine until the orders double, and a plan priced by executions moves with them. Ask what the next tier costs before you need it.
- Retries and polling. A trigger that checks every minute runs far more often than the job it is watching, and a step that retries on failure can multiply a quiet month into an expensive one. Both are settings, and both are worth reading before launch.
- Somebody's time. Self-hosting is not free, it is paid in hours: upgrades, backups, and the evening it stops. If those hours are yours, price them at what your own time is worth, then compare that with a hosted plan honestly.
What makes the price go up?
These are the things I count when I quote an n8n job. Each one is real work, and each one is also something you can choose to leave out of the first version.
- How many systems it touches, and whether each one has a usable interface or needs a workaround.
- What happens when a step fails: retry, alert, hold for a person, or roll back. The failure path is most of the difference between a demo and something you can run a business on.
- Whether a model reads or writes anything, which adds a usage bill and a testing burden that a plain data move does not have.
- Who has to approve the output. Anything a customer reads stays behind a person until the accuracy earns otherwise, and that is a daily cost, not a build cost.
- Whether it has to be handed over: documentation, a walkthrough, and workflows you can edit yourself rather than ones only I can read.
When should you hire someone, and when should you not?
I would rather tell you this before you spend anything.
- Do it yourself when the job is two systems, both with a clean connector, and you have an afternoon. The tool is learnable and the first workflow is a good way to find out whether the process is even worth automating.
- Hire someone when other people depend on the output, when a failure costs you a customer or a payment, or when the process crosses systems that do not talk to each other cleanly.
- Hire someone when it already exists and stopped working. Taking over a half-finished setup is usually cheaper than rebuilding it, and it is work I do regularly.
- Do not hire anyone yet when the process itself changes every week. Automating a moving process means paying twice, once for the build and once for the rebuild.
What should a quote include before you agree to it?
Whatever the arrangement, these five lines are what make two quotes comparable. If a quote does not have them, you are comparing prices for different products.
- The workflows, named, with what triggers each one.
- The failure behaviour for each: what it does when a step fails, and who finds out.
- What is explicitly not included, so the gap is a decision rather than a surprise.
- Where it runs and whose accounts it lives in, because that decides what you keep if the work ends.
- What the first month after launch looks like, and what it costs.
What to remember
- A single workflow starts at $49 and a build around one process starts at $499, as one fixed price; ongoing care starts at $149 a month.
- Fixed price for a job you can write down, a small paid piece of work to produce the plan when you cannot, and a monthly arrangement for anything that must keep running.
- Budget twelve months, not one. Volume, retries and somebody's time are what move an n8n bill after launch.
- A workflow that fails quietly costs you twice: you keep paying for it and you stop getting the work it was meant to do.
Sources
Primary documentation, for anyone who wants to check the details.