Guide · Automation templates and integrations

Can AI read n8n, Zapier and Make template and integration pages?

Updated 2026-11-12 · kagentx

Short answer

Tested 2026-11-12: none of the robots.txt files of n8n, Zapier and Make blocks template or app pages, and Zapier even names and allows the main AI crawlers. The difference is in page content: Zapier's template and app pages carry their text in the HTML, and the template pages use question-style headings; n8n's app page is very complete, but its template detail page has only a title, node names and one truncated sentence in the HTML; our curl was stopped by a challenge page on Make's list and app pages, so what real crawlers see is unverified. When building integrations and templates, fill in names, descriptions and titles to each platform's official rules, and write a separate integration page on your own site. Whether AI cites these pages we have no evidence, and it is unverified.

Are the n8n, Zapier and Make pages open to crawlers?

None of the three robots.txt files blocks template or app pages, but they are written very differently. On 2026-11-12 we read each platform's robots.txt and sitemap with curl:

Checkn8nZapierMake
robots.txt200, 76 bytes; a single * group that only disallows /cdn-cgi/; no AI crawler named200, 1,862 bytes; the * group disallows /app-directory/, /apps/*/landing/, /zapbook/, /api/ and every URL with parameters; a separate group names GPTBot, OAI-SearchBot, ChatGPT-User, ClaudeBot, PerplexityBot and others, all Allow: /; a Content-Signal line at the end200, 499 bytes; the * group has Allow: / and disallows /integration/*, */template/*, /integrations/*/*/* and URLs with search parameters; no AI crawler named
sitemap12 sub-sitemaps; 14,148 workflow URLs, 2,677 integration URLs, and 39,052 in the first of the app-pair filesrobots.txt lists 11 sitemaps; how many URLs each holds is unverified6 sub-sitemaps; the templates sitemap has 7,830 URLs in the form /templates/<number>-<English phrase>
AI-specific entry pointsllms.txt exists and lists Markdown versions of reports and other contentrobots.txt has an Agentmap line pointing to .well-known/ai-catalog.jsonNot checked

These are URL counts, not counts of indexed pages; how many pages exist outside these sitemaps is unverified.

What is in the HTML of the template and app pages?

The difference is how much a crawler can read without running scripts. For each platform we fetched a list page, a template detail page and an app page (2026-11-12, curl; visible text is the character count after removing scripts and styles, including header and footer menus, so only the order of magnitude matters):

PageStatus and sizetitle and meta descriptionStructured dataVisible text
n8n.io/workflows/ (template list)200, 544 KBPresent; self-referencing canonicalNoneAbout 1,300 characters: the heading is there, the list of template cards is not in the HTML
n8n template detail page (no. 1459)200, 83 KBtitle ends with "| n8n workflow template"; the meta description is a truncated sentence1 WebApplicationAbout 1,100 characters: title, node names, author, category; no h2
n8n.io/integrations/slack/ (app page)200, 377 KBPresent; self-referencing canonicalWebApplication plus FAQPage (5 questions)About 9,800 characters; 12 h2s, including Supported actions and FAQ
zapier.com/templates (template list)200, 3.5 MBPresent; self-referencing canonical1 SoftwareApplicationAbout 31,000 characters
Zapier template detail page (the Slack + ClickUp one)200, 664 KBtitle names both apps and "template"; self-referencing canonicalNoneAbout 11,600 characters; 10 h2s, mostly questions, including Overview and FAQ
zapier.com/apps/slack/integrations (app page)200, 1.0 MBPresent; self-referencing canonical6: Organization, BreadcrumbList, FAQPage, SoftwareApplication and two ServiceAbout 28,700 characters; 24 h2s
Make template detail page (no. 11359)200, 173 KBtitle ends with "| Workflow Template | Make"; has a meta description and a self-referencing canonicalNoneAbout 5,400 characters; h2s include About template and Similar templates
make.com/en/integrations/slack and /en/templates403, about 5.6 KBBoth are a "Just a moment..." challenge page with noindex,nofollowNoneAbout 60 characters; asks to run scripts

How to read this: Zapier's template and app pages carry their text in the HTML, and the template detail page puts purpose, method and audience under question-style headings, which is the form AI needs for citation. n8n's template detail page has only a title, node names and one sentence, and the real steps live in the workflow canvas, which a crawler that does not run scripts cannot read; n8n's app page, by contrast, is very complete. Our curl was stopped by a challenge page on Make's list and app pages; we have not verified whether real search crawlers are stopped the same way. Whether crawlers run scripts is covered in Do AI crawlers render JavaScript?.

One number worth noting: on n8n's template detail page and app page the aggregateRating is identical, 4.6 from 27 ratings. We have not verified whether the rating is each page's own data. Rating markup must match the reviews actually visible on the page; see Does structured data help AI search? for how to write it.

Which fields do you fill in to turn a product into an integration?

All three platforms start with the integration (app) and then the template. Below are only the rules we could find in the official docs (read 2026-11-12; none of the pages shows a date):

ItemZapier (integration)n8n (community node)Make (custom app)
NameMust match the core brand exactly, without TM, "app" or "integration"Package name starts with n8n-nodes- or @scope/n8n-nodes-Must use a service Make does not have yet
DescriptionThe short description starts with "<name> is a…", has no links, does not mention Zapier or other apps, sells less and explains what the app doeskeywords in package.json must include n8n-community-node-package; the docs do not specify the other fields one by oneEvery module needs a correct label and description
Homepage URLThe marketing website, not the app or a login pageNot verifiedNot verified
LanguageAll user-facing text is English onlyNot verifiedNot verified
Triggers and actionsBuilt on the platform as triggers, actions and searchesNodes and credentials go in the n8n attribute of package.jsonNeeds a universal module; search and trigger modules need a limit parameter, and pagination if the API supports it
ReviewReviewed by Zapier after submissionTo get n8n verification, from 2026-05-01 you must publish with GitHub Actions and a provenance statementThe Make QA team reviews the code and requires every module to appear in at least one test scenario
{
  "name": "n8n-nodes-example-notes",
  "keywords": ["n8n-community-node-package"],
  "n8n": {
    "nodes": ["dist/nodes/ExampleNotes/ExampleNotes.node.js"],
    "credentials": ["dist/credentials/ExampleNotesApi.credentials.js"]
  }
}

The above is an example of an n8n community node package.json; the package name and paths are placeholders we made up. The name and description end up in the app page's title and meta description, so write what the product does and who it is for in the first sentence.

And what do you fill in to turn it into a template?

A template is a ready-made flow for "these two apps doing this one thing", and it is the landing page for queries like "X to Y automation". The three platforms' rules differ a lot:

ItemZapier (Zap template)n8n (workflow template)Make (template)
Who writes the title and descriptionYou doGenerated automatically by the Creator portal; you do not write themSubmission rules unverified
TitleStart with a verb, sentence case, present tense; say new or updated; plural items; no sync or automaticGenerated automaticallyNot verified
Description2-4 sentences, starting from the user's problem or goal; no Zap, trigger or action jargon; no links; extra setup tips start with Note:; each one must be uniqueGenerated automatically; but the canvas needs one overview sticky note of 100-300 words with the headings How it works, Setup and CustomizationNot verified
Other requirementsOnly public integrations can be used; Paths, Code, Webhooks, Looping and Formatter steps are not allowed for nowNodes need descriptive names; workflows with four or more nodes need section notes; no secrets or personal data; export it and test the import in a fresh workflowNot verified
ReviewUsually within 2 weeksNew creators can submit one at a time; after three are approved they become verified creators and can submit up to four at onceNot verified
Title:        Create Example Notes entries for new Typeform responses
Description:  Want every form response saved as a note without copying it by hand?
              This automation adds a new note in Example Notes each time Typeform
              receives a response. Note: Choose the notebook during setup.

This is a title and description written to Zapier's rules; the app names are placeholders. Zapier states that descriptions that only swap the app names are rejected, and n8n lists batches of near-identical workflows as unacceptable, so do not generate templates in bulk from one pattern.

How do you make integrations and templates easier to find, step by step?

  1. 1. Build the integration first, then the template

    Zapier templates can only use public integrations, and n8n templates that use community nodes must name the package and the tested version. Without a listed integration, a template has nowhere to land.

  2. 2. Write the first line of the description as a one-sentence definition

    "<name> is a…": what it does and who it is for. Zapier's rules even specify the sentence pattern; n8n and Make do not, but the same pattern works there too.

  3. 3. Put both apps and the action in the template title

    Following Zapier's rules: the trigger app first, "new" or "updated" stated, plural items. This structure is also how buyers phrase "X to Y" searches.

  4. 4. Write the n8n overview note as a product description

    100-300 words saying who the workflow is for, what it produces, which accounts and credentials it needs, and what its limits are. Because the template page's HTML does not carry the canvas, the sticky note is the closest thing to body text.

  5. 5. Also write an integration page on your own site

    Store and directory page formats are set by the platform and you cannot change them. On this one page of your own site, write the supported actions, triggers, setup steps and limits, then link to each platform's page. See How do you write an "X and Y integration" page so AI will cite it? for how.

  6. 6. Use question-style headings and an FAQ

    Zapier's template detail pages do exactly this: answer first, then expand. See How to write content AI will quote: before and after and Is FAQ structured data (FAQPage) still useful? for how.

  7. 7. Re-test a few weeks later with the same questions

    Ask ChatGPT and Perplexity "is there an n8n or Zapier template that connects X to Y", ask each question 3 times, and note whether you are mentioned and which URLs are cited.

Similar directories compared: How do you get Notion templates, Figma Community files and Framer templates recommended by AI?, How do Slack apps and Discord bots get recommended by AI?, How do you get an API product recommended by AI?, How do you get an MCP server recommended by AI and listed in connector directories?.

FAQ

Can AI read n8n's template pages?

Tested 2026-11-12: the template detail page's HTML has the title, node names, author, category and one truncated sentence in the meta and JSON-LD; there are no h2s and no step descriptions. The workflow itself lives in the canvas, which a crawler that does not run scripts cannot read. Whether AI cites it is unverified.

Are Zapier's template and app pages served as HTML text?

In our test, yes. The template detail page has about 11,600 characters of visible text and 10 h2s; the app page has about 28,700 characters, with FAQPage and SoftwareApplication structured data. We did not imitate an AI crawler; we used a desktop browser User-Agent.

Can crawlers read Make's pages?

Our curl got 200 and the full text for a Make template detail page, but 403 for the integration page and the template list, which returned a "Just a moment..." challenge page with noindex. That is the result from our network exit; whether real search crawlers are challenged the same way is unverified.

What is Zapier's robots.txt attitude to AI crawlers?

In the robots.txt we read, GPTBot, OAI-SearchBot, ChatGPT-User, ClaudeBot, Claude-User, Claude-SearchBot, Google-Extended, PerplexityBot and DuckAssistBot share one group with Allow: /; the file ends with Content-Signal: ai-train=yes, search=yes, ai-input=yes. This is the content on 2026-11-12 and it can change.

What hard rules apply to Zap template titles and descriptions?

Titles start with a verb, use sentence case and present tense, say new or updated, use plural items, and avoid sync and automatic; descriptions are 2-4 sentences, have no links, avoid Zap, trigger and action jargon, and cannot just swap the app names. Source: Zapier's official docs, read 2026-11-12.

Do I have to write the title and description of an n8n template myself?

No. The official docs state that the Creator portal generates the title and description. What you write is the overview sticky note on the canvas (100-300 words), the section notes and descriptive node names.

What does a community node have to satisfy?

The package name starts with n8n-nodes- or @scope/n8n-nodes-, keywords include n8n-community-node-package, and nodes and credentials go in the n8n attribute of package.json; to get n8n verification, from 2026-05-01 you must publish with GitHub Actions and a provenance statement.

How do you submit a template to Make?

Unverified. We only found Make's review rules for custom apps and the public API reference for templates, and no public document on template submission fields.

Sources

kagentx

Can crawlers read your integration page?
Let the evaluation tell you

The evaluation checks whether crawlers can get in and whether the text is served directly, then asks ChatGPT, Perplexity and Claude (with web search on) your buyers' questions and sees whether you are mentioned. The Basic plan is $1, and results usually take 2–3 minutes.