Osum · Updated
Make.com vs n8n vs Power Automate (2026)
Pick an automation tool and you inherit its failure story: what happens at 2 a.m. when a field is empty. Make.com vs n8n vs Power Automate is really a choice between a per-credit canvas that builds fastest, a self-hostable builder that gives up control slowly, and a Microsoft stack that follows your IT department. Here is how the three compare on cost, control and what happens when a run breaks.
The short answer
Pick Make if you want working automations this week on a visual canvas and can accept per-credit billing. Pick n8n if you want unlimited steps per execution and can host it or learn a more technical builder. Pick Power Automate if your company already runs Microsoft 365 and IT owns the automations. Pick Zapier if a pre-built template covers the whole job. Whichever you pick, the skill that decides whether it saves you time is designing the route before you build it.
Solo builders report Make is the fastest to learn and the easiest to debug; the same builders describe scenarios that break on real data and credit bills that surprise them. Both problems have the same fix, and it is a design habit, not a tool swap.
What changed in 2026
Make now bills in credits, not operations. On standard apps one operation still equals one credit, so old habits survive; the change matters because AI modules can bill tokens on top, with usage tags on the module itself. A scenario that was cheap with plain modules can jump the day you add an AI step, and Make’s own pricing FAQ publishes the token rates.
Make also added Maia, a prompt-based co-worker that builds and fixes scenarios, and the Make AI Agents app for agent steps inside flows. Maia is in public beta on paid plans, with a 30-day trial on Free accounts; the AI Agents app is in open beta on all plans. Both change what beginners can build fast. Neither changes what survives a busy Tuesday: that is still the route you design.
What this changes for the pick: tool comparisons written before the credits change quote per-operation prices that no longer exist. Judge Make on credits per run plus tokens for AI steps, and check the pricing page the week you decide.
Make.com and Integromat are the same tool
Make was called Integromat until early 2022, when the company renamed the product Make and moved to make.com. Older forum threads, videos and course listings still say Integromat. Same canvas, same scenarios, same company; if a tutorial predates the rename, expect an older interface around the same ideas.
Two words trip beginners: a scenario is Make’s word for one automated workflow, and a module is one step inside it. Rival tools say workflow and node or step. The concepts transfer; the vocabulary does not.
Make, n8n, Power Automate and the rest
Three things decide this pick: where it runs, what a run costs, and who notices when a run fails. Make runs in Make’s cloud and bills per credit. n8n runs hosted or self-hosted and bills per workflow execution, so complex flows cost the same as simple ones. Power Automate runs in Microsoft’s cloud, priced per user or per bot, and answers to your IT department. Everything else on this page is a variation on those three answers.
| Option | Where it runs | Paid from (checked Sep 28, 2026) | Best at | Watch out for |
|---|---|---|---|---|
| Make | Make’s cloud | Free plan; $9 a month for 10,000 credits | Fast visual builds with branching, filters and error handlers | Per-credit billing adds up with AI steps and high volume |
| n8n | n8n Cloud or self-hosted | Cloud from €20 a month billed annually; free self-hosted Community Edition | Unlimited steps per execution, code nodes, full control | A more technical builder, and self-hosting is your upkeep |
| Power Automate | Microsoft cloud | Premium $15 per user a month, paid yearly | Deep Microsoft 365 and enterprise governance fit | Premium connectors and desktop RPA sit on higher tiers |
| Zapier | Zapier’s cloud | Free plan; Professional from $19.99 a month | Simple app-to-app jobs from a huge template shelf | Task-based pricing gets expensive on multi-step jobs |
Make.com vs n8n: who runs the machine
n8n gives you more of the machine. Every integration is included, one execution can hold unlimited steps, the code node drops real JavaScript or Python into a flow, and the self-hosted Community Edition runs on your own server for nothing. Builders who hit Make’s per-credit wall on high-volume AI work describe the switch as economics, not preference.
Make gives you more of your week back. The canvas is faster to learn, branching and filters are visual, error handlers attach to the module that fails, and Make runs the servers, the updates and the uptime. Beginners report building multi-path flows in a week; the same week on self-hosted n8n often goes to Docker, tunnels and node versions.
Choose n8n if volume makes per-credit billing hurt, you want code inside flows, or your data must stay on your server. Choose Make if you want the working automation this month and would rather pay credits than run infrastructure.
Make.com vs Power Automate: buy the tool or the suite
Power Automate’s real product is fit. If your company runs Microsoft 365, flows sit next to your mail, your files and your compliance rules, with governance your IT team already owns. Desktop RPA on higher tiers drives legacy software no canvas app can reach, and process mining maps the work before you automate it.
Make’s real product is reach and speed outside one vendor’s wall. More than 2,000 app integrations connect in an afternoon, the canvas teaches itself, and a solo operator never opens an admin center. Per-user pricing that looks cheap for ten Microsoft flows looks different for a freelancer serving five clients on five stacks.
Choose Power Automate if Microsoft owns your company’s data story and someone else pays the license. Choose Make if the apps you actually use decide, and you are the one paying per seat.
What Make costs right now
Make’s plans price by credits per month. Each module action in a scenario counts as one credit on standard apps, so a three-module scenario processing one lead costs about three credits. AI modules can bill tokens on top; Make’s own AI provider publishes token-per-credit rates by model, and modules with dynamic usage carry tags such as Tokens, File size or Run time.
If you run out of credits, you can upgrade, buy extra credits, or turn on auto-purchasing. Auto-purchasing is convenience with an open tab: cap it or budget first.
Every price below was checked on the official Make pricing page on September 28, 2026.
| Make plan | Monthly price | Credits | Notable limits |
|---|---|---|---|
| Free | $0 | Up to 1,000 a month | 15-minute minimum schedule interval; no time limit |
| Core | $9 | 10,000 a month | Unlimited active scenarios; scheduling down to the minute |
| Pro | $16 | 10,000 a month | Priority execution; custom variables; full-text log search |
| Teams | $29 | 10,000 a month | Teams and roles; shared scenario templates |
Where Make scenarios break
A market-stall owner built the classic first scenario: an email trigger, a filter skipped, a sheet write. It worked all weekend. On Monday the inbox held three spam rows, one renamed sheet had broken the mapping, and a connection error had stopped the run mid-flow with a lead inside it. Nothing was on fire. Work was simply piling up where nobody looked.
The pattern repeats across public posts. One builder: “Was banging my head against the wall with make.” Another: “Webhooks not working. The first run works, I get an accepted notice, then it doesn’t stay connected in the scenarios.” A third: “was having trouble with make, troubleshooting more than 6 hours for one scenario, and i made a switch to n8n. resolved in 30 mins.” The tool was rarely the whole story; the route was undesigned.
Each pattern has a fix you can build in minutes. Filter before the paid modules, so junk stops for the price of a check. Turn on incomplete executions, so a failed run waits in a tab with its error instead of vanishing. Read the history tab weekly, because a green tick can carry warnings. And price the run before you add an AI step, because tokens bill on top of credits: “while you can predict the no. of credits used in a scenario , you won’t be able to track the token usage that accurately.”
The lesson 1 brief: log one lead without copy-paste
This is the job lesson 1 of Osum’s Make course puts in your hands, copied word for word. Build it in about 15 minutes on the Free plan, on your own sheet, with nothing sent to anyone else.
Build a scenario named Lead Desk. Add a Custom webhook trigger and name the hook lead-in. Add a Google Sheets Add a Row module on a sheet with the columns Name, Email, Note, and map each field to its own column. Then email one test lead to the lead-in address and run the scenario once.
- The webhook started the scenario when your email arrived.
- The row holds three values in three columns, nothing merged.
- The run bubble shows the operations, so you know what one lead costs.
- Nothing left your own accounts: you emailed yourself, to Make.
Make.com FAQ
Is Make.com free? The Free plan costs $0 with up to 1,000 credits a month and no time limit. Scheduled scenarios wait at least 15 minutes between runs on Free, and the Core plan at $9 a month lifts that to minute-level scheduling with 10,000 credits.
What is a Make credit? One module action in a scenario is one credit on standard apps, so a lead through a two-module scenario costs about two credits. AI modules can bill tokens on top, and Make publishes the rates in its credits guide.
Make.com vs n8n: which is cheaper? It depends on volume and steps. Make bills per module action, so complex flows bill more; n8n Cloud bills per workflow execution with unlimited steps, from €20 a month billed annually, and its self-hosted Community Edition is free but runs on your server. Price your own busiest week both ways before you decide.
Make.com vs Power Automate: which should a small team pick? Pick Power Automate when your work lives in Microsoft 365 and IT owns governance, at $15 per user a month paid yearly for Premium. Pick Make when your apps span vendors and you want a visual builder you can learn in a week.
Can Make run scenarios without me? Yes. A webhook scenario starts the moment data arrives, and scheduled scenarios run on their own while active. A stopped scenario is the catch: incomplete executions store failed runs only after you turn the setting on.
Do my scenarios keep running if one step fails? No, not by default: the run stops and, on scheduled triggers, Make deactivates a scenario after repeated errors. With Store incomplete executions on, the failed run waits in a tab with its data and error for a retry.
Is this course made by Make? No. Osum is independent and not affiliated with Make or the apps you connect, and the certificate is an Osum certificate.
Verdict: which automation tool should you pick?
Make is the pick for solo operators and freelancers who want branching automations running this week without servers: the fastest canvas to learn, filters and error handlers built in, from $9 a month. Choose n8n when volume or code demands it and someone can host. Choose Power Automate when Microsoft already owns the stack. Choose Zapier when a template is the whole job.
Whichever you pick, the skill that decides whether it saves you time is the same: designing the route before you build, so junk stops early, failures wait for a fix and every run costs what you expected. We call it the Fail-safe Scenario. If you’re going with Make, our course teaches it in 12 hands-on lessons, one working lead desk from first row to scheduled system, with a certificate and every other Osum course included.