Make vs n8n: Which Should You Pick?

The meter: credits against executions

Make's billing unit is the credit. Its help centre says credits replaced operations as the term for the billing unit, that existing plan pricing is unchanged, and that by default one operation equals one credit. Its pricing FAQ says most actions cost one credit, and that reading, searching, creating, updating, deleting, transforming, aggregating and iterating data all count. The Router module and the error handler modules (Rollback, Break, Resume, Commit and Ignore) are stated not to use credits. A single run can cost from 2 credits to thousands depending on what it does.

n8n bills one workflow execution per run. Its Starter plan is 20 euros a month billed annually for 2,500 executions, and Pro is 50 euros a month billed annually for 10,000.

So a ten module scenario that runs a thousand times a month uses roughly 10,000 Make credits, while the same logic is 1,000 n8n executions. Make is therefore closer to a per step meter than to n8n's per run meter, with free routing and error handling as a partial offset. A polling trigger that finds nothing still spends credits; a Make help centre guide gives the trigger alone at about 288 credits a day when polling every five minutes, though that page is old and the figure should be checked.

Make's prices are in US dollars and n8n's are in euros, so the two should not be compared number against number. Make's own comparison page says paid plans start at 9 US dollars a month for 10,000 credits and that its free tier gives 1,000 credits a month. A help centre page gives 113.85 US dollars for 150,000 credits a month on Core, and another gives 338.13 US dollars a month for 500,000 credits.

Make's plans and the catches

Make names its plans Free, Core, Pro, Teams and Enterprise, with a separate Start-up offer of 50 percent off Business for organisations under 20 employees. Business is not otherwise described on the pricing page. Paid plans scale by credit tier rather than by fixed bundles. Core is documented as covering up to 300,000 credits a month, with higher usage moved to Pro at the same monthly cost, and Pro as covering up to 8 million. The full per plan feature table did not load for this research, so plan by plan feature limits are not stated here.

Several catches are documented. Credits expire at the end of the term, and extra purchased credits expire one month after purchase on monthly plans or at year end on annual Pro and Teams plans. Extra credits cost 25 percent more than credits included in a plan, according to Make's notice of plan adjustments dated from November 6, 2025. When credits run out, scenarios stop until more are added, though incoming webhooks queue up to a storage limit and polling scenarios resume from their last successful run.

Make's pages also disagree with each other on size. The pricing FAQ says Standard Apps are more than 1,000 third party apps, its homepage says 3,500 or more apps in one place and 3,000 or more native apps in another, and its comparison page says both over 3,000 and 3,500 or more. n8n's integrations page lists 2,366 integrations. Treat Make's count as somewhere above 1,000 and up to 3,500 depending on which page and which definition of app you read, and check the specific apps you need.

Hosting and licence

Make is SaaS only. Its comparison page describes it as cloud first at all subscription levels, with Make handling servers, updates and uptime, and none of the pages read describe a self hosted edition. Its homepage lists GDPR compliance, SOC 2 Type II and SOC 3, encryption in transit and at rest, SSO, role based access and audit logs, but does not say where data is hosted.

n8n publishes a Community Edition on GitHub that can be self hosted under the Sustainable Use License. That licence is not OSI approved open source. It permits internal business use and non commercial use, and bars selling a product or service whose value comes substantially from n8n functionality. n8n's Business plan is self hosted only at present, and self hosted Business and Enterprise instances need a licence key that contacts n8n's licence server daily. The Starter and Pro plans are hosted by n8n.

The practical split is simple. If data must stay on your own infrastructure, Make is not an option and n8n is. If you would rather never patch or back up a server, Make removes that work, and so do n8n's hosted plans, though those carry execution based limits.

Error handling, scheduling and the builder

Make's editor is a visual scenario canvas where modules are wired together, and its homepage says scenarios can be built by prompting, by drag and drop or through MCP, with custom JavaScript or Python available for developers. Its error handling is built from error handlers attached to modules, using the directives Ignore, Resume, Commit, Rollback and Break. Its help centre also links an exponential backoff article, which describes automatic retries with progressively longer delays. The help centre pages for error handler routes and incomplete executions did not load, so how each directive behaves is not described here beyond its name.

Scheduling offers eight options: fixed interval, run once, daily, weekdays only, weekly, monthly, specific dates and on demand. Daily runs can have multiple times. Make's help centre says the minimum interval depends on the plan and links to the pricing page, which did not load, so the per plan minimum is not confirmed here. Check it before relying on short polling intervals.

n8n is a node graph aimed at engineers. It has JavaScript and Python code nodes, custom nodes on self hosted instances, and workflow JSON that suits version control. Both tools handle branching and loops; the difference is audience. Make's canvas is designed so that someone who does not write code can follow a scenario; n8n expects the builder to be comfortable reading API documentation and handling data structures.

AI agents

Make's AI Agents app describes an agent as an autonomous system with a model, a system prompt, long term context such as a knowledge base, tools and visible reasoning. Tools can be modules, whole scenarios and MCP tools. The help centre page read says the app was in open beta, with functionality and pricing subject to change. Make's notice dated from November 6, 2025 says all paid plans can connect their own AI provider, which was previously limited to Pro and higher. Some AI features cost more than one credit, depending on factors such as tokens, file size, page count or processing time, and the pages read give no fixed figure.

n8n puts an AI Agent node, MCP Server Trigger and MCP Client directly into the workflow graph, and its self hosted Assistant uses your own API key with no credit limit. The two approaches differ in where they spend: Make's agents draw on the credit meter, with variable cost for model heavy steps, while n8n's agent runs inside one execution. Agent quality needs testing on your own tasks.