Make vs n8n: Cost Math with Credits and Executions, Hosting, License and AI

Most comparisons stop at "Make for no-code, n8n for developers." That is not enough to make a business decision. What matters more is how each tool counts usage, where your data lives and what you are allowed to do when you host it yourself. All prices and quotes below come from the vendors' own pages, as of October 6, 2026.

What is the key difference between Make and n8n?

The key difference is the billing unit. Make counts every single step, n8n counts the whole run.

• Make bills in credits. According to its pricing page, each module action, such as adding a row in Google Sheets or fetching Gmail data, counts as one credit. A module is a building block in a scenario (Make's name for a workflow). According to the Make Help Center, credits replaced the earlier unit "operations"; for regular apps, one operation still equals one credit.
• n8n bills in executions. According to its pricing page, an execution is a single run of the entire workflow, no matter how many steps it has or how much data it processes.

That gives you a simple rule of thumb for Make: credits per month ≈ runs × (1 for the trigger + modules after the trigger × records per run). The trigger is the starting block that kicks off a run, and it counts once per run. On n8n, only the runs count.

How much does n8n cost, and how much does Make cost in 2026?

Both start in the low double digits, but for very different volumes.

Sources: Make pricing, n8n pricing. The Make prices are starting prices from the pricing page as of October 6, 2026, depending on billing cycle. Make offers larger credit packages through a slider. The page only shows their price once you pick one, so work out your volume first and then check.

How do I convert my workflow?

Count modules and records per run, and the answer is clear. Two typical cases:

Case 1: Lead from the website form into the CRM. The workflow runs once per inquiry: receive the form, validate the data, enrich the company, create the contact in the CRM, notify sales, send a confirmation. That is six modules. At 1,500 inquiries a month:

• Make: 1,500 × 6 = 9,000 credits, fits into Core with 10,000 credits
• n8n: 1,500 executions, fits into Starter with 2,500

Make is cheaper here under either billing cycle: even if the $9 already includes the annual discount, Core paid monthly still stays well below the €20 for n8n Starter.

Case 2: Nightly sync. Once a night, a workflow reads 200 items from the ERP system and sends each one through four modules: convert, validate, update in the online store, log.

• Make: 200 × 4 = 800 credits per night plus the trigger; over 30 nights, roughly 24,000 credits a month plus triggers. That is more than double the 10,000 package.
• n8n: 30 executions a month, barely making a dent in Starter

n8n clearly wins here, and the gap grows with every record. If you build a sync in Make that quietly grows from 200 to 2,000 records, you only notice it in the credit counter, and in the worst case when your quota runs out.

What do AI steps cost in Make and n8n?

AI steps make the bill variable on Make. According to the Make Help Center on credits:

• If you use Make's own AI provider, credits are based on tokens (the text units a model reads and writes) and operations.
• If you connect your own account with OpenAI, Anthropic (Claude) or Google (Gemini), you only pay Make for the operations. The AI provider bills the tokens separately.

On n8n, an AI step stays inside one execution. You pay the model costs through your API key with the provider. According to the pricing page, Starter also includes 1,600 credits for the built-in assistant.

Our take: for AI-heavy scenarios in Make, use your own API connection. That way you see model costs separately at the provider, and your credit math stays predictable.

Where is my data stored?

With n8n Cloud, it sits in Frankfurt. With Make, it sits in the EU or North America.

• n8n states on its pricing page that data from hosted plans is stored in the EU, on servers in Frankfurt, Germany.
• Make lists its hosting as "AWS (EU/North America)". Check which region your organization is in and document it in your record of processing activities under the GDPR.
• You can also run n8n entirely on your own infrastructure, on your own server or with a German hosting provider. Make is cloud only.

Keep in mind: hosting location alone does not make a workflow GDPR compliant. A data processing agreement, data minimization and the connected services are part of it too. This is not legal advice; when in doubt, ask your data protection officer.

Can I self-host n8n for free?

Yes, for your own internal purposes. Not as a paid service for others. The Community Edition is released under the Sustainable Use License. It allows use and modification only for your own internal business purposes or for non-commercial and personal use. You may only distribute the software free of charge and for non-commercial purposes. Files with ".ee." in their name are not covered at all and require an enterprise license.

What does that mean in practice?

• Allowed: You run n8n on your server and automate your own processes, such as invoice filing, lead routing or reports.
• Check first: A service provider sets up workflows in your own n8n instance. In our understanding, that remains your internal use, because the instance belongs to you.
• Not covered: A provider hosts n8n and sells access to it, or builds it into a paid product for customers. That requires an agreement with n8n.

And free does not mean zero cost: servers, updates, backups and monitoring take time. If nobody on your team will look after it, the cloud version is the better choice.

When should you pick Make, and when n8n?

The answer depends on three questions: how many records per run, who maintains it, and where the data has to live?

Our take: for midsize companies, one tool for everything is rarely the right answer. We often build lead pipelines in Make, because marketing can read and adjust them there, and heavy data syncs in n8n. What matters is that every process lives in exactly one place and is never duplicated.

What does this mean for you?

Before you sign up for a plan, run the numbers with your real data:

1. List your three most important workflows.
2. For each one, write down runs per month, modules per run and records per run.
3. Calculate Make: runs × (1 for the trigger + modules × records). Calculate n8n: just the runs.
4. Add a 30 to 50 percent buffer, because data volumes grow.
5. Decide where the data may live and who will maintain the system.

Using HubSpot? Then first read what is changing with HubSpot private apps, because it affects Make and n8n alike. If you would rather do the math with us: we build AI automation with Make and n8n and tell you upfront which tool is cheaper for your process. Get in touch.