NewCapture the real searches ChatGPT runs

See the searches behind AI answers.

ChatGPT, Gemini and Google's AI Mode don't search for your query. They fan it out into a dozen searches and write the answer from what comes back. queryfanout shows you those searches: simulate them on Gemini and OpenAI side by side, capture the real ones ChatGPT runs, and search 4 fan-out queries from everyone's runs.

Works withGeminiOpenAIChatGPT
Fan-out · live
best electric SUV for snowEV winter range test resultsAWD electric SUV reviewsRivian R1S vs Model Y snowEV ground clearance comparedheat pump EVs cold weatherYOUR PROMPTBest EV fordriving in snow?AIanswer
Fan-outs run
0
plus 2 ChatGPT conversations captured
Simulation → real hit rate
—
run with real-search passes to measure this
Gemini ∩ OpenAI overlap
—
run both models side by side to measure this
Sub-queries indexed
4
4 real searches · 2 prompts
Run

Fan out a query

Simulate the fan-out on Gemini and OpenAI, then run real-search passes to see which searches each model actually runs.

New fan-out
Pick a provider to see the projected cost.

Every pass is a live call on your own API key. Real-search passes answer with the provider's web search and record the searches it actually ran (search tools are billed per query).

How it works

From one question to the searches behind the answer

Read the guide: how query fan-out works →
  1. 01

    Enter a query

    Type the question a customer would ask. Pick Gemini, OpenAI or both.

  2. 02

    We run live fan-outs

    Each model breaks the query into the sub-queries it would search, over several passes, on your own API keys. Counts show which sub-queries are stable.

  3. 03

    Compare against reality

    See where the models agree, how the fan-out clusters by topic, and which searches each model actually runs when it answers with the web.

  4. 04

    Learn from everyone's runs

    Every run, plus real ChatGPT searches captured by the Chrome extension, feeds an open index you can search by keyword or by meaning.

Why we built it

AI search doesn't rank your query. It rewrites it.

ChatGPT, Gemini, Google AI Mode and AI Overviews answer by running many searches on your behalf and writing the answer from what those searches retrieve. Ranking for the head term isn't enough anymore: your content has to be retrieved for the sub-queries, and those sub-queries don't show up in any rank tracker.

We built queryfanout at iPullRank, after our free Qforia simulator, so SEOs, content teams and researchers can see the fan-out, measure how stable it is, check it against the searches models really run, and plan content against it. The open index already holds 4 fan-out queries.

  • 01
    Visible

    The hidden searches behind AI answers, simulated and captured.

  • 02
    Measurable

    Frequency across passes, overlap between models, and how well the simulation predicts real searches.

  • 03
    Open

    Every run is public and searchable, so the whole industry learns from it.

From iPullRank

Is your brand ready for AI search?

Take the AI Readiness quiz to see where you stand and what to prioritize first.

Take the AI Readiness quiz →
The index

Search everyone's fan-outs

0 runs · 2 seed queries · 4 real searches captured · Capture ChatGPT's searches with the extension →

Recent runs

Everything run on queryfanout.com is public and feeds the index.

No runs yet. Be the first.