Visual automation platform
Make.com
A visual scenario builder with a very large connector library. We ran three controlled tests on the free plan.
Have we tested this?
We tested it Yes. See the workflow pages for run logs. Free plan, zone us2, tested 2026-09-22. Three scenarios against our own synthetic endpoint.
How it bills you
Billing unit: Credits (the unit changed from 'operations' — older tutorials describing operations are out of date)
Billing units are not interchangeable between platforms. A “credit”, a “task” and an “execution” measure different things, and comparing the headline prices without comparing the units is how people get surprised.
Free tier: Free plan. We tested on it and hit no paywall in any of our three scenarios.
| Plan | Monthly | Annual (per month) |
|---|---|---|
| Free | $0 (as of 2026-09-22) | — |
| Core | $10.59 (as of 2026-09-22) | $9 |
| Pro | $18.82 (as of 2026-09-22) | $16 |
| Teams | $34.12 (as of 2026-09-22) | $29 |
What we can say in its favour
We tested itIn our tested 10-line workflow a completely failing run cost 4 credits against 13 for the identical successful run — Make stopped at the first failing item rather than attempting the rest.
We tested itIn our test, three automatic in-run retries of a failing HTTP step were billed as a single operation.
We tested itIts error panel named the origin, showed our destination's own response body, and offered to build an error-handler route for us.
We tested itSearching 'HTTP' in the app found the HTTP module immediately, and it accepted a nested line-item path directly.
Limitations
We tested itIn our tested 10-line workflow a successful fan-out cost 13 credits: 1 trigger + 1 fetch + 1 iterator + 10 per-line writes. Cost scaled with line items, which your customers control.
We tested itMake bills the trigger and the iterator as well as the external calls. Every module in the scenario counted.
We tested itA run that failed on a duplicate cost the same as a successful run, and a configuration error that never left Make still cost 4 credits.
We tested itNo automatic re-execution occurred after our deterministic failure, and the Incomplete Executions queue was empty — that storage is off by default. Recovery was manual.
We tested itNo alert reached anyone by default. We did not test whether alerting can be configured.
⚠ We could not verify thisWe measured a 10-line order but never a 1-line order, so Make's per-line slope rests on one data point plus its own operation counter.
Who should not do this
Anyone whose orders routinely carry many line items and who needs a predictable monthly bill — in our test the cost scaled one credit per line. Also anyone who needs unattended recovery from a failed run: in our test Make did not retry on its own.
What we earn if you use this
A commission, if you sign up through our link. We joined Make.com’s affiliate programme on 2026-09-24. Its published rate is 35% of subscription payments only — not on purchases of extra operations, 12 months, and the clock starts at registration rather than first payment. We have earned nothing from it so far. When that changes, this line changes.
AXORD may earn a commission if you create a Make.com account through the link below. It does not change what you pay — the link opens Make.com’s ordinary registration page at Make.com’s published prices — and it does not change our testing, our measurements or our conclusions. We still recommend free and non-affiliate options where they are the better answer.
Make.com bills in Credits (the unit changed from 'operations' — older tutorials describing operations are out of date). Prices above were checked on 2026-09-22. How this unit compares with the other platforms — the three units do not convert.
Free ways to do this first, and the Decision Sheet for working out whether the task is worth automating at all.