What questions are AI engines actually answering for your topic — and who are they citing?
Type a topic. The tool harvests questions from Google PAA, Reddit, and Stack Exchange, clusters them semantically with Claude, then shows you the AI Overview presence flag, search volume, KD, and the actual Perplexity answer + cited sources for the top clusters. Free, ~30 seconds, cached searches are unlimited.
- Harvesting questions from Google PAA, Reddit, Stack Exchange
- Clustering questions semantically with Claude
- Pulling search volume + KD + AI Overview flags
- Querying Perplexity for top clusters
- Done — rendering
Question clusters
Top 5 clusters include Perplexity answer + cited sources. Click a cluster to expand.
What this tool actually does
The tool runs the full 7-step AI search keyword research workflow on a topic in about 30 seconds. Step 1 — discovery: harvests questions from Google's "people also ask" via DataForSEO, top Reddit threads on your topic, and Stack Exchange (for technical topics). Step 2 — clustering: groups the 50-100 raw questions into 6-10 semantic clusters via Claude. Step 3 — scoring: pulls search volume, keyword difficulty, and AI Overview presence from DataForSEO. Step 4 — engine benchmark: queries Perplexity for the top 5 clusters and returns the actual generated answer plus cited source URLs.
What it does not do (yet): the v1 ships without the citation gap matrix and content brief generation. Both are in v2 — the citation gap matrix surfaces "your competitors are cited X times in AI answers, you're cited 0 times" per cluster, and content briefs translate clusters into ready-to-write outlines with passage structure, schema type, and citation-worthy data points. If you want either now as a paid engagement, the AI search optimisation service page describes the full programme.
How the tool keeps it free
Aggressive caching. Most topic queries are repeated by other users; the first search costs us about $0.10 to compute, every subsequent search of the same topic for 28 days costs $0. About 80% of queries hit cache after the first month of use. Plus a soft IP rate limit of 10 fresh searches per day — far above what an honest researcher needs in a day, low enough that scrapers cannot drain the budget.
If the rate limit ever bites, that is a signal we are popular enough to fund a paid tier. Until then, free.