Is Competitor Intelligence free?
Yes. The entire report, all eleven sections, is free with no email gate and no signup required. Anonymous users get 3 runs per tool per rolling 24 hours; a free account raises that to 10.
How long does a report take?
Typically one to two minutes. Most of that is live research: resolving each company, scraping up to four pages per company, running four categories of market signal searches per company, and reading your review URL. The analysis itself is a single large model call.
Where does the data come from?
Three places: the companies' own websites (homepage, about, pricing, and features pages, scraped live), recent web coverage from the past year (funding, hiring, launches, and positioning searches, listed as sources in your report), and the review page you optionally provide. The model analyzes that evidence rather than answering from memory, and claims that lean on category knowledge instead of evidence are labeled with lower confidence.
How accurate is the pricing intelligence?
Pricing scraped from a public pricing page is real and current as of the run. When a competitor hides pricing behind a sales call, the tiers shown are estimates inferred from market positioning and are labeled with lower confidence. Use estimated tiers to prepare for the conversation, not as quotable numbers.
Why does a competitor's section look thin or generic?
Usually the research could not reach their pages: the site blocked scraping, the pages do not exist, or a bare company name resolved to the wrong domain. The research confidence score on your report reflects this. Re-running with the competitor's exact website URL fixes the most common cause.
Do I need to provide a review URL?
No, but it is the single biggest quality upgrade. With a review page (a G2, Capterra, or Trustpilot listing for a competitor), the review sentiment section is mined from real customer positives, complaints, and churn signals. Without one, that section is generated from category knowledge and marked low confidence.
Will two runs with the same inputs give the same report?
Within 24 hours, yes, exactly: identical inputs return the same stored report without using a run. After that window a fresh run re-researches everything, so the report reflects whatever changed on the competitors' sites and in the news since.
What happens if part of the run fails?
You get a partial report: the sections that completed render normally and a retry button re-runs the whole pipeline after a 30 second cooldown. If nothing completed, the run is marked failed and you can start again. Failed runs are never served from cache.
Can competitors tell I researched them?
The research fetches their public pages the way any crawler or browser would; it does not log in, fill forms, or probe anything non-public. At most, a standard page request appears in their web analytics.