Watch AI fan a topic into its real questions.

Straight answer

The hidden sub-queries decide whether you get cited.

A real fan-out in three steps

1. 01Add

   Your own API key

   Paste a free Google AI Studio (Gemini) key — and optionally a Perplexity key for a second engine. The keys stay in your browser: requests go straight to Google and Perplexity, never to a GEO Toolbox server.

2. 02Run

   A real fan-out, live

   Enter a topic. We call each engine directly and read the actual sub-queries it fired while answering — Gemini's grounded searches, Perplexity's related questions. Real engine output, not an LLM guessing.

3. 03Read

   Queries, intents, divergence

   See every query tagged Fired or Related, clustered into intents — and, with two engines, the cross-engine divergence map: the intents both engines share versus the whitespace only one explores.

• 01Is it really free, and do you see my API key?

  Yes, and no. The demo is completely free and needs no sign-up. Your API key is used entirely in your own browser — every request goes directly from your browser to Google's (and Perplexity's) official endpoint. The key is never sent to, proxied through, logged by, or stored on a GEO Toolbox server. Reload the page and it's gone. You only pay for whatever those engines charge your key, which for a fan-out is a fraction of a cent.
• 02Where do I get a Gemini API key?

  Free, in two minutes, at aistudio.google.com — sign in with a Google account, click “Get API key”, and copy it. Google's free tier is generous enough to run plenty of fan-outs. Paste it into the field above; we never see it.
• 03Why only Gemini and Perplexity — where are ChatGPT and Grok?

  Browser security (CORS) decides this, not us. Google's and Perplexity's APIs allow direct browser calls; OpenAI's does not, so ChatGPT can't run from a free in-browser demo. Grok could, but the demo keeps to the two that give the cleanest fan-out. The full in-app feature runs all four engines (it calls them from our server), plus the parts a browser simply can't do — real search volume, page-coverage checks, and the citation landscape.
• 04How is the demo different from the paid feature?

  Same core idea, a slice of the depth. The demo shows the real fan-out and, with two engines, the divergence map. The in-app AI Query Fan-Out adds ChatGPT and Grok, real search volume (honestly labelled), a check of how well a URL you give it already covers each intent, the citation landscape (who AI cites for these queries), and a ranked content worklist — metered per scan on our keys.
• 05I ran it and got no queries — why?

  Gemini only fans a topic out when it decides to ground its answer in a web search. Very broad or very personal prompts sometimes get answered from the model directly, with no sub-queries to show. Try a more search-like, commercial topic — “best CRM for startups”, “enterprise password manager”, “running shoes for flat feet” — and you'll see the fan-out.