Integration3 min read2026-07-31

AI visibility in Make

A Make scenario with three modules — a schedule, an HTTP request, a router — is a complete AI-visibility monitor. AgentGEO supplies the part Make can't: the raw answers ChatGPT, Perplexity, Gemini, Google AI Overviews, Google AI Mode and Copilot actually give, with every cited source as structured data. There's no AgentGEO app in Make's directory; the integration is the built-in HTTP module against one REST endpoint. That keeps it honest and portable — and it means the one thing to get right is the timeout, which this page covers rather than hides.

Read this page with an AI

Make's iterator and aggregator modules are unusually good at the core GEO chore: walking a run's sources[] array, filtering for your domain or a competitor's, and aggregating the result into one Slack message or Sheet row.

Connect AgentGEO to Make

Add an HTTP → Make a request module: method POST, URL https://api.agentgeo.org/v1/fetches, an Authorization: Bearer ag_live_... header, body type JSON:

HTTP module — request content
{
  "query": "best crm for startups",
  "surfaces": ["chatgpt", "perplexity"]
}

Set the module's timeout as high as your Make plan allows and keep the surface list short per call — the API holds the request while live surfaces answer, up to 180 seconds for a slow one. Two fast scenarios beat one that times out. Each response record carries surfaceKey, status, the verbatim answerText, sources[] (title, URL, position), latencyMs and providerRecordId.

If your scenarios keep brushing the timeout, flip the pattern: create a schedule in the AgentGEO console (or POST /v1/schedules) so the slow fetching happens server-side, and have Make poll GET /v1/runs — a fast, credit-free read — then pull full answers with GET /v1/runs/{id}. The Zapier page walks the same poll pattern step by step; it maps one-to-one onto Make modules.

Scenarios worth building

  • Six-engine citation diff. Fetch the same query across all six surfaces, iterate every sources[], aggregate to a table of who-is-cited-where, post to Slack weekly.
  • Category watch to a data store. Daily fetch of your category query → append records to a Make data store → alert only when the cited set changes from yesterday.
  • Client reporting feed. For agencies: per-client scenarios that append each run to the client's Sheet, feeding the white-label GEO report you assemble monthly.
  • Content follow-up. A week after publishing, fetch the target query and record whether the new page shows up in sources[] — a closed feedback loop on GEO work.

More integrations

Same contract, different runners: n8n for self-hosted control, Zapier for the poll pattern, Pipedream when you'd rather write the step in code. See all integrations.

FAQ

Direct answers to the questions this page raises.

No — the integration is Make's built-in HTTP module against POST https://api.agentgeo.org/v1/fetches. Nothing to install, and the wiring is portable to any other platform with an HTTP module.

Raise the module timeout, request fewer surfaces per call, or move the slow part out of Make entirely: a server-side AgentGEO schedule does the fetching, and your scenario polls GET /v1/runs — a read that returns in moments regardless of how slow the underlying scrape was.

Run the HTTP response through an Iterator on the record's sources[] array. Each bundle then carries one citation's title, URL and position — ready for filters ("URL contains ourdomain.com") and aggregators.

One delivered record = one credit; failed records cost zero, and reads (GET /v1/runs) are free. Usage bills by consumption under a spend cap. Build against an ag_test_ key first — labelled demo records, zero credits; live ag_live_ keys are minted in the console and come with a plan.

Keep reading

Where this page leads next.

Run these checks on your own brand

Two ways in. Send a URL and a person runs the fetches for you, free — or connect your agent over MCP, on a plan, and run them yourself. Either way you get the raw answers and their citations, never a score.