n8n vs Make vs Zapier: how to choose in 2026

All three tools connect apps and move data between them. The real differences are in what they charge for, who can build on them comfortably, and where your data runs. We build on all three at SBD, and pick per client, so this guide sets out the trade-offs as we see them rather than arguing for one.

Prices below were checked on 29 October 2026 and change often. Treat them as a starting point and check the live pricing pages before you commit.

At a glance

	Zapier	Make	n8n
Billing unit	Task: each successful action step	Credit: most module actions, some AI features use more	Execution: one full workflow run, however many steps
Free tier	100 tasks a month, two-step Zaps only	1,000 credits a month	No free cloud plan; free trial, and the self-hosted Community Edition
Entry paid plan	Professional from $19.99 a month billed annually, 750 tasks	Core, 10,000 credits (price: see note below)	Starter €20 a month billed annually, 2,500 executions
Integrations	9,000+ apps	3,000+ native apps	Fewer native nodes; HTTP requests and code cover the rest
Self-hosting	No	No	Yes, Community Edition under the Sustainable Use License
Hosting region	Zapier cloud	Make cloud	n8n cloud stores data in Frankfurt, or anywhere you self-host
What does not count	Triggers, polling, Formatter, Filter, Paths, failed actions	Error handler and Router modules	Steps within a run
Best suited to	Simple automations built by business users	Visual, branching scenarios at moderate cost	Multi-step, technical or data-sensitive workflows

A note on Make prices: Make’s pricing page did not show plan prices in a form we could read when we checked. Third-party summaries from September 2026 report Core at about $9 a month billed annually for 10,000 credits. We use that figure below but flag it as less certain; check the slider on Make’s page.

Pricing at three example volumes

The same workflow is counted differently on each platform. Take a typical workflow with one trigger and five action steps:

• Zapier: the trigger is free, each of the five actions is a task, so 5 tasks per run.
• Make: each module run uses a credit, including the trigger module, so roughly 6 credits per run.
• n8n: 1 execution per run.

Volume	Runs a month	Zapier tasks	Make credits	n8n executions
Small	300	300 × 5 = 1,500	300 × 6 = 1,800	300
Medium	1,500	1,500 × 5 = 7,500	1,500 × 6 = 9,000	1,500
Large	8,000	8,000 × 5 = 40,000	8,000 × 6 = 48,000	8,000

Priced at published annual rates (Zapier Professional, Make Core, n8n cloud):

Volume	Zapier	Make	n8n cloud
Small	1,500 tasks is above the 750 tier; the 2,000 task tier is $49 a month	1,800 fits in 10,000 credits, about $9 a month	300 fits in Starter (2,500), €20 a month
Medium	7,500 tasks needs the 10,000 tier, $129 a month	9,000 fits in 10,000 credits, about $9 a month	1,500 fits in Starter, €20 a month
Large	40,000 tasks needs the 50,000 tier, $289 a month	48,000 credits needs a higher credit tier or extra bundles; price it on Make’s slider	8,000 fits in Pro (10,000), €50 a month

What this shows:

• Step count drives Zapier and Make bills, not n8n cloud. Double the steps and the Zapier and Make columns double; n8n stays the same.
• Make is often cheapest at low and medium volume for the same number of steps, if the reported price holds.
• Zapier’s Team plan costs more for the same tasks ($69 for 2,000, $169 for 10,000, $399 for 50,000 billed annually) and is what you need for shared workspaces and more than one seat.
• Monthly billing costs more. Zapier states annual billing is 33% off the monthly rate.
• AI steps and agents can be metered separately. On Zapier, AI and Code steps count as tasks and some actions cost more than one; Make says some AI features use more than one credit. See the Zapier tax in 2026 for a worked bill including AI steps.

When to pick each

Zapier suits teams where people across the business build their own small automations. The builder is the easiest of the three, the app catalogue is the largest, and Filter, Formatter and Paths do not use tasks. It gets expensive as workflows get longer or busier.

Make suits teams who want a visual canvas for scenarios with routers, iterators and error handlers, without writing code. Routers and error handlers do not use credits, which helps for branching logic. It is cloud only.

n8n suits teams with some technical confidence: you can drop into JavaScript or Python, call any API, and run AI agent workflows with your own model keys. It is the only one of the three you can self-host, and its cloud plans store data in the EU. Whether self-hosting is worth it for a small team is a separate question, covered in n8n self-hosted or cloud for a team under 50.

If you are weighing an AI agent instead of any of them, read does an AI agent still need Zapier? first. For most businesses the answer is that agents and workflow tools do different jobs.

More detail on each platform: Zapier, Make and n8n.

1. List the workflows and count the steps

Write down every workflow you expect to run in the next year, not just the first one. For each, note:

• what triggers it,
• how many times it runs a month (use real numbers from the source system where you can),
• how many action steps it has, including any AI steps.

This list is the input for every number that follows. Without it, you are comparing headline prices that measure different things.

2. Convert volume into each billing unit

Using the method above, convert your list into Zapier tasks, Make credits and n8n executions. Then price each at the smallest published tier that covers it, with headroom for growth. Note the date you checked, and price AI and agent usage separately where a platform meters it separately.

Our free build vs buy tool helps frame the wider choice, and the ROI calculator shows whether the workflows pay for themselves on any of the three.

3. Check the hard constraints

Some factors rule a platform out regardless of price:

Constraint	Question to answer
Data residency	Must data stay in a region? n8n cloud stores data in Frankfurt; self-hosting puts it where you choose. Check Zapier’s and Make’s current hosting terms against your requirement.
Self-hosting	Do you need to run it on your own infrastructure? Only n8n offers this.
Licence	If self-hosting n8n, the Sustainable Use License allows internal business use but not reselling n8n as a service. Some enterprise features need a paid licence key.
Apps	Is every app you need a native integration? If not, can you reach it by HTTP request or webhook?
Builders	Who will build and maintain workflows: business users, an operations team or developers?

4. Pilot two real workflows

Build two workflows on your top two platforms: one simple and one awkward, with branching, a loop over records and a step that can fail. Compare:

• Time to build and how readable the result is to someone else.
• Error handling. What happens when a step fails: can you retry, route the failure, and get alerted? Zapier does not charge for failed actions; Make has error handler modules that do not use credits; n8n lets you set a dedicated error workflow.
• Actual usage after a week, against your estimate from step 2.

5. Decide, document and set a review date

Pick the platform and write down why, with the numbers. Name an owner for the account and for failure alerts. Set a review date, and revisit sooner if volume doubles or you add AI steps at scale.

Migration notes

If you are moving between platforms:

• There is no direct import between them. Each workflow is rebuilt, so start with an inventory and migrate the busiest or most expensive workflows first.
• Run old and new in parallel for a short period, writing to a test destination or with the old one paused per workflow, so nothing is processed twice.
• Re-create credentials in the new platform; they do not transfer.
• Check trigger behaviour. Polling intervals and webhook handling differ, so a workflow that ran every few minutes may behave differently after the move.
• Recheck usage after a month against the estimate, then cancel the old plan.