Contents
At a glance
- Unit
- One CPU second of workflow processing time
- Not billed
- Step count - Latenode states unlimited steps do not multiply the bill
- Free allowance
- 10,000 CPU seconds per month (checked August 2026)
- First paid tier
- $0.00012 per CPU second from 10,001 to 100,000 (checked August 2026)
- Second tier
- $0.00011 per CPU second from 100,001 to 200,000
- Highest volume tier
- $0.00005 per CPU second above 5,000,000 per month
- Biggest cost driver
- Waiting - slow APIs, backoff delays and long model calls bill while blocked
The unit is time, not credits
People search for "Latenode execution credits" and the honest answer is that Latenode does not use them. Latenode bills in CPU seconds: the processing time your workflows consume while they run. Latenode states this directly as its differentiator, noting that unlimited workflow steps do not multiply your bill.
This is a third billing model, distinct from both the per-action family (Zapier tasks, Make operations) and the per-run family (n8n executions). Adding filters, branches or helper steps does not increase cost proportionally. Only actual execution duration matters.
The published rates
As listed on Latenode's pricing page in August 2026, the Free plan includes 10,000 CPU seconds every month, which Latenode describes as supporting roughly 10,000 executions if each one runs about a second. Beyond the monthly free allowance, pay-as-you-go rates are tiered by volume: 10,001 to 100,000 CPU seconds at $0.00012 per second; 100,001 to 200,000 at $0.00011 per second; with rates decreasing further at higher volumes, down to $0.00005 per second above 5,000,000 CPU seconds a month.
Ten thousand CPU seconds is under three hours of compute. That is the number to hold on to: the free tier is generous for fast workflows and evaporates on slow ones.
What makes the count spike
Waiting. Any step that sits blocked - a slow third-party API, a retry with a backoff delay, a model call that takes eight seconds to return a long answer - is consuming the meter while it does nothing. On a per-action platform that same slow call costs one task regardless of whether it takes 200 milliseconds or twenty seconds. On Latenode the difference is a hundredfold.
The design consequences invert the usual advice. Parallelise where you can. Set aggressive timeouts, because a hung request bills until it gives up. Prefer a webhook to a poll, because polling wakes the workflow to do nothing. And treat a slow upstream integration as a cost problem, not just a latency problem - see polling vs webhook and retry and backoff.
What you should know
CPU-second billing makes Latenode very cheap for the workflow shape that is expensive everywhere else: many steps, heavy loops, large data transforms, all completing fast. It makes Latenode expensive for the shape that is cheap elsewhere: a small number of steps that each wait on something slow.
Model this properly before comparing. Take one real workflow, time it, and multiply. A workflow that runs in 0.4 seconds, 50,000 times a month, is 20,000 CPU seconds - 10,000 free plus 10,000 billed at $0.00012, about $1.20. The same 50,000 runs on a per-task platform with two actions each would be 100,000 tasks, which is a different order of magnitude of money entirely.
Latenode CPU second FAQ
Does Latenode use execution credits?
No. Latenode bills in CPU seconds - the processing time your workflows use while running - rather than per execution, per step or per credit. This is the main thing to correct if you are comparing it against Zapier, Make or n8n.
How many CPU seconds does Latenode include free?
10,000 CPU seconds every month as listed in August 2026, which Latenode describes as supporting roughly 10,000 executions if each takes about a second. That is under three hours of compute in total.
How much does Latenode cost per CPU second?
Tiered by volume beyond the free allowance: $0.00012 per second from 10,001 to 100,000 CPU seconds, $0.00011 from 100,001 to 200,000, and decreasing further at higher volumes to $0.00005 above 5,000,000 a month.
Do extra steps cost more on Latenode?
Not directly. Latenode states that unlimited workflow steps do not multiply your bill - adding filters, branches or helper steps does not increase cost proportionally. Only how long the workflow takes to run affects the meter.
Why is my Latenode bill high when my workflow is simple?
Because time, not complexity, is the meter. A three-step workflow where one step waits eight seconds on a slow API bills eight seconds every run. Set tight timeouts, parallelise independent steps, and replace polling with webhooks so the workflow is not awake doing nothing.
When is CPU-second billing cheaper?
When workflows are complex but fast: many steps, big loops, heavy transforms, all completing quickly. It is more expensive than per-run billing when a small workflow spends most of its life blocked on slow external calls.
Primary source
Latenode - pricing and CPU-second billing
Checked August 2026. Anything priced or rate-limited moves, so verify against the source before you budget.
Want this built rather than explained?
Knowing what latenode cpu second means is one thing; having it running in your business is another. We scope one workflow at a time, quote a fixed price in writing after a free consult, build it, and hand it over with documentation you own.
