MEASURING n8n.
Plenty of opinions circulate about how n8n workflows are built and very few numbers. These are the numbers, taken from public sources with the method written down, so they can be checked rather than believed.
the figures, measured 2026-07-28
- 12
- The median published n8n template has 12 nodes
- 55.5%
- 55.5% of published n8n templates contain at least one AI node
- 91.1%
- 91.1% of published n8n templates carry at least one sticky note
- 14%
- Only 14% of published n8n templates configure error handling on any node
- 54.1%
- The most-used node type in published n8n templates is n8n-nodes-base.code, in 54.1% of them
- 60,401
- Published n8n templates contain 60,401 sticky notes between them
A census of all 10,957 public n8n.io templates — no sampling error. Public n8n.io templates only. Templates are built to be shared, so they skew simpler than production workflows and over-represent whatever the library is promoting — currently AI. This describes a published n8n template, not a private production workflow.
The anatomy of a published n8n template
How big a template really is, how it starts, and what is in it — node-count distribution, trigger mix, AI adoption, error-handling rates and the most-used integrations, across every one of the 11,015 public n8n.io templates. CSV and JSON included.
Muhammad Muneeb. “The anatomy of a published n8n template.” Pixtex, 2026-07-28. Census of all 10,957 public n8n.io workflow templates, measured 2026-07-28. CC BY 4.0. https://pixtex.dev/research/anatomy-of-a-published-n8n-template
CC BY 4.0. Quote it, chart it, build on it — attribution to the report page is all we ask.
Using these numbers
Quote them, chart them, build on them — attribution to Pixtex — The anatomy of a published n8n template is all we ask, and every copy button on these pages attaches it for you. Every figure states what it was measured over and how. Where we measured everything, we say so; where we sampled, the sample is seeded so anyone can redraw the same set from the same public API. If there is a cut of the data we have not published, ask and we will run it.
Measured and published by Muhammad Muneeb. The rules behind every classification are on the report page under how these are counted, and the comparator ranks your own workflow against the library without uploading it.
Guides, not just numbers
The measurements say what people build. The guides say how to get a picture of it out of n8n and into a README, a client deck or a template listing.