How to List Your Product on Zapier, Make, and n8n So AI Agents Can Find It

What does listing do for AI agents?

Agents choose tools by reading their names and descriptions, so an integration that exists but is poorly described may still be skipped.

How do you list on Zapier?

Build the integration in Zapier's Platform UI or CLI, test it with invited users, then click Submit for Review. Zapier's publishing requirements include:

• A publicly launched product, not invite-only or in beta.
• Clear, current API documentation covering the endpoints used, linked in the submission.
• Ownership or permission. Only the company that owns the API, or a developer it contracts, can publish a public integration.
• English for all user-facing text.
• A non-expiring test account using Zapier's integration-testing email address.
• A factual description that starts with your app's name, in the form “YourApp is a…”, explaining what it does rather than selling it.

Zapier says a reviewer replies within a week. Approved integrations appear in the app directory with a Beta tag for 90 days; embedding Zapier in your product can end the beta early once a sign-up comes through it. Zap templates, help articles and in-product links then drive usage.

How do you list on Make?

Make calls these custom apps. You build the app in Make's developer tools, remove test modules and connections, click Publish, and choose which modules to make visible. The review request asks for a link to the API documentation and to scenarios that use your modules. Review has three stages: a form about you and the API, an automatic review, and a manual review by Make's QA team. Once approved, the app is available to all Make users. Publishing cannot be undone, and published modules cannot be deleted, so finish testing first.

How do you list on n8n?

n8n integrations are community nodes, distributed as npm packages whose names start with n8n-nodes-. Anyone can publish one, but verified nodes are the ones users can discover and install from the nodes panel inside n8n. To verify:

1. Build with n8n's node CLI so the node follows the expected structure, and pass the linter and local tests.
2. Publish to npm through GitHub Actions with a provenance statement. Since May 1, 2026, n8n does not accept verified nodes published from a local machine.
3. Submit through the Creator Portal for automated checks and n8n's review.

For agent use, mark the node as usable as a tool, so n8n's AI Agent node can call it.

What makes an integration easy for agents to use?

An agent decides which action to run from its name, description and fields. The same things that help human users help agents:

• Name actions as verbs and objects, such as “Create invoice” or “Find customer by email”.
• Describe each action in one plain sentence, including what it returns.
• Offer search actions, so an agent can look up a record before updating it.
• Keep required fields to a minimum, with sensible defaults.
• Return clear error messages that say what to fix.
• Keep API documentation accurate, because agents and reviewers both rely on it.

Product descriptions that match buyers' words are covered in how to write product descriptions AI can match to buyer queries.

Do directory pages help AI search visibility?

They can. Zapier's robots.txt explicitly allows GPTBot, OAI-SearchBot, ChatGPT-User, ClaudeBot, Claude-SearchBot, PerplexityBot and others, and declares that content may be used for AI training, search and AI input. Its app and integration pages are therefore open to AI search crawlers answering “does X integrate with Y?”. n8n's robots.txt is open too. Make's robots.txt allows its app pages and app-pair pages, such as /en/integrations/slack, and blocks only deeper paths and an older /en/integration/ path, so its main directory pages are also open to crawlers.

Directory pages should back up your own. Keep an integrations page on your site listing each connection and what it does, as covered in GEO best practices for SaaS companies. Your own documentation can also be made agent-readable directly, as covered in how llms.txt and MCP servers make documentation machine-readable.

Which platform should you start with?

• Zapier if your buyers are business users. It has the largest catalogue.
• n8n if your buyers are technical teams or AI agent builders, many of whom self-host.
• Make if your buyers run detailed operations workflows and already use it.

Each listing is ongoing work: APIs change, reviews recur for updates, and users expect support. Start with one, measure connections and usage, then add the next.

We have not published an integration on these platforms ourselves. This page is based on Zapier's, Make's and n8n's developer documentation and their robots.txt files.

What might this look like in practice?

Consider an illustrative invoicing tool for freelancers, Tallysheet, with a documented REST API. It published a Zapier integration with six actions, including “Find client by email” and “Create invoice”, each with a one-sentence description. After the review and a month in beta, users of Claude and ChatGPT with Zapier MCP began asking their assistants to “invoice this client in Tallysheet”, and the agent found the actions without setup beyond connecting an account. Six months later the team added a verified n8n node for its more technical users.

What do people ask about listing on automation platforms?

Can AI agents use my product if it is on Zapier?

Yes, through Zapier MCP, for users who connect your app to their Zapier account.

Can I publish an integration for someone else's API?

Not publicly on Zapier unless you own the API or the owner has hired or permitted you.

How long does Zapier's review take?

Zapier says a reviewer replies within a week; approved apps then spend 90 days in beta.

Does an n8n node need verification?

Not to publish, but only verified nodes can be installed from the nodes panel inside n8n.