Enter a search query and the page you want to rank. We show you the real sub-queries Google Gemini and Claude run, who they cite, and what your page needs to become the answer.
Trusted by teams at
Starting AI engines
Three steps to see what AI engines actually search for — and how to win the citation.
Type the keyword you want to rank for and the page you're optimizing. Any query, any URL.
Real API calls to Claude + Gemini fire sub-queries and return what they actually searched — not simulated.
See what's missing from your page with prioritized, writer-ready actions to fix it.
Currently analyzing: Best Procurement Software. Every report shows real backend queries, cited competitors, coverage gaps, and a prioritized action plan your team can execute today.
Try it yourself — it's freeBuilt for teams that want to rank in AI-powered search, not just traditional SERPs. See what no other tool shows you.
We fire actual API calls to Claude and Gemini. No keyword databases, no guesswork — real observed sub-queries.
Gemini uses the same Google Search grounding as AI Mode. The closest observable proxy to what AI Mode actually does.
Every report includes writer-ready action items with HIGH/MEDIUM/LOW priority. Hand it straight to your content team.
10+ years in SaaS SEO across 70+ B2B brands. We know what actually moves rankings in AI search.
Firing queries across AI engines...
This runs real API calls to Claude + Gemini. Takes ~30-60 seconds.
Query fan-out is the process AI search engines use when answering a question. Instead of running a single search, they decompose your query into multiple sub-queries — reformulations, comparisons, entity lookups, recency checks — and fire them all in parallel before synthesizing a final answer. This tool shows you those real sub-queries.
We make real API calls to Claude (Anthropic) and Gemini (Google) with your seed query. Claude uses its web_search tool — every sub-query it decides to run is captured from the API response. Gemini uses Google Search grounding and returns its actual webSearchQueries in the response metadata. Nothing is simulated.
Yes. Every query you see in the results is a real search that Claude or Gemini actually ran against a live search index. Gemini's grounding uses the same Google Search infrastructure as AI Mode, making it the closest observable proxy to what Google AI Mode does behind the scenes.
Traditional keyword tools show you what humans search for. This tool shows you what AI engines search for when a human asks them a question. AI engines fan out a single user query into 4-10+ sub-queries that are completely different from what appears in any keyword database.
We recommend running an analysis before creating or updating any high-priority content piece, and re-running monthly for your top 10-20 keywords to catch shifts in which sub-queries the engines fire and which competitors they cite.
No. Queries are processed in real-time and not stored. Your target URL is fetched once for analysis and immediately discarded. We don't log queries, URLs, or results.