n8n vs Make Pricing in 2026: How Each Counts Usage, With Worked Examples

The n8n vs Make pricing question comes down to what each tool counts. n8n counts one execution for every complete run of a workflow, however many steps it contains. Make counts credits (the new name for operations) for the steps inside each run, so a scenario that loops over many items spends far more of its allowance than the same job does in n8n.

That difference matters more than the headline plan prices. Two tools with similar monthly fees can differ several times over once a workflow handles lists of records. The sections below explain how each vendor meters usage, what the plan tiers look like as of November 2026, and a method you can repeat with your own workflows. Pricing and limits last verified 6 November 2026 on the vendors' own pages; both change, so check the pages linked in each section before you decide.

How does n8n count usage?

n8n's pricing page defines it plainly: an execution is a single run of your entire workflow. The page states that the number of steps and the amount of data processed do not change the count, and that pricing is based on monthly workflow executions regardless of complexity. A workflow that fires once, loops over 500 rows, calls an API 500 times and writes 500 results is still one execution.

Two things sit alongside that count. First, concurrency: each plan caps how many executions can run at the same moment. Second, duration and retention limits, such as how long one run may last and how many past executions n8n keeps for you to inspect. These limits, not the execution count, are usually what a heavy user hits first on the cheaper hosted plans.

How does Make count usage?

Make bills in credits. Its help centre says credits replaced operations as the name for the billing unit, and that by default one operation equals one credit for non-AI apps. Only features triggered by scenario runs, such as app modules and some in-app features, spend credits.

The detail that surprises people is per-bundle processing. Make's help article on operations explains that a module processes each bundle (each item of data) separately, so each bundle triggers its own module run. Its own example is a form trigger that returns 10 bundles passing through three further modules: one operation for the trigger check plus ten for each of the other three modules, 31 in total. A trigger counts once per check, regardless of how many bundles it returns.

AI features are metered differently again. Make's help page says some AI and advanced modules vary by tokens, file size, pages processed or run time, and that using Make's own AI provider is billed to you in credits, while a custom AI connection means you pay Make for operations and your model provider for tokens. Anything AI-heavy needs its own estimate rather than a flat one-credit-per-step assumption.

What do the plans look like in November 2026?

n8n

n8n's page shows euro prices billed annually (the page I read showed annual prices only):

All plans are described as including unlimited users and workflows, and the self-hostable Community Edition is free. n8n also lists a 50% discount on Business for eligible companies under 20 employees. If a hosted plan's execution quota runs out, the page lists capped overage fees on Starter and Pro; for Business and Enterprise it says workflows keep running and overage charges may apply if you do not move up a tier. Read the n8n pricing page for the current numbers, because tiers and the Business plan's hosting options have shifted before.

Make

Make has a Free plan with 1,000 credits per month, paid plans called Core, Pro and Teams, and a custom Enterprise plan. On the paid plans you choose a monthly credit volume, and the price follows that choice. Extra credits are sold in bundles of 1,000 or 10,000, and auto-purchase of 10,000-credit bundles is available on Core, Pro and Teams. Credits expire at the end of the term, and bought extra credits expire after one month on monthly plans.

The usage allowance also scales with credits: per 10,000 credits per month, Make's page lists 5 GB of data transfer and 10 MB of data storage. I could not read Make's per-plan prices from the page when I fetched it, because that part loads dynamically and did not come through as text, so this article quotes no Make dollar figures. Open the Make pricing page, set the credit slider to the monthly total from the worked example below, and compare it with the n8n tier.

How do you estimate cost at your volume?

Use this method for any workflow. Count the steps that run each time, multiply by the number of items processed per run, multiply by the number of runs per month. That gives Make credits. For n8n, ignore steps and items and just count runs. Then compare each total with the plan tiers.

Here are three illustrative workflows. The volumes are invented for the arithmetic; replace them with your own.

1. Lead intake. A form webhook fires per new lead, then three steps: enrich the company, create a CRM record, post to chat. Four modules in Make. Assume 200 leads a month. Make: 4 × 200 = 800 credits. n8n: 200 executions.
2. Nightly order export. One scheduled run each day collects the day's 500 new orders and writes each to a spreadsheet, with three modules processing every order. Make: 1 trigger credit plus 500 × 3 = 1,501 credits a run, times 30 days = 45,030 credits. n8n: one execution a day, 30 a month.
3. Support ticket triage. A webhook fires per ticket, then five steps including one AI model call. Six modules in Make. Assume 3,000 tickets a month. Make: 6 × 3,000 = 18,000 credits, before any extra token-based credits for the AI step. n8n: 3,000 executions, plus the model provider's token charges in both tools if you bring your own key.

On those numbers Make needs roughly 20 times as many billing units as n8n needs executions. The 3,230 executions would not fit n8n Starter's 2,500 but sit comfortably inside Pro's 10,000. For Make you would price a volume that covers about 64,000 credits a month. Notice where the load comes from: the nightly export, the only one of the three that loops, produces over 70% of Make's total from a single scheduled job.

Flip the example and the gap closes. If every workflow ran once per event with no loops and few steps, the counts would sit within a small multiple of each other, and Make's credit price per unit could win. The method tells you which situation you are in.

When does n8n win, and when does Make win?

n8n tends to win when

• Workflows loop over lists, batches or large result sets, since the loop costs no extra executions.
• Workflows have many steps, branches or sub-workflows, because step count is not billed.
• You want the option to self-host the same workflows later; the self-hosted versus cloud cost breakdown covers that decision.
• You need custom code, API calls to unusual services, or AI agent logic that would otherwise spend credits at every step.

Make tends to win when

• Workflows are short and run per event, so credits per run stay in single digits.
• Non-technical staff will build and maintain scenarios, and Make's visual builder and large app catalogue save more time than the credit count costs.
• You want a managed service with no server decisions at all and your volume is modest. Its free tier gives 1,000 credits a month for testing.
• You would rather pay by a volume slider than think about execution tiers and concurrency limits.

What else changes the real bill?

Four details move the total more than the sticker price. Polling triggers: Make counts a trigger once per check, so a scenario that polls every few minutes can spend credits even on quiet days, and a webhook trigger avoids that; confirm in your account how empty checks are counted. Error handling: retries and fallback routes are extra module runs. Run time: n8n Starter stops a run at five minutes, which breaks long batch jobs that Pro's 40-minute limit allows. People: a cheaper tool that one person must babysit costs more than a dearer one that runs untouched.

For the wider tool question beyond price, including Zapier, see our comparison of n8n and Zapier for business automation. If you would rather have the workflows designed and metered for you, the n8n workflow automation service page describes how we build them, and the cost estimator gives a first range.

Frequently asked questions

Is n8n cheaper than Make?

For workflows that loop over many items or have many steps, usually yes, because n8n bills per run and Make bills per step per item. For short, once-per-event workflows the two can be close. Run the three-step arithmetic above with your own numbers before deciding.

Does Make still charge by operations?

Make now calls its billing unit a credit. Its help centre says credits replaced operations as the term, and that for non-AI apps one operation equals one credit by default. Some AI modules use more.

What counts as one n8n execution?

One complete run of a workflow, from trigger to finish, according to n8n's pricing page. The number of nodes and the amount of data do not change the count.

Can I avoid the limits by self-hosting n8n?

The free Community Edition has no execution quota, but you then pay for the server, backups and upkeep, and some paid features are not included. The licence also sets rules on commercial use. Our self-hosting cost guide goes through them.

Which is easier for a beginner?

Many people find Make's visual scenario builder easier to start with, while n8n rewards people comfortable with data structures and a little code. Price should be second to who will maintain the workflow.