Support

Trust Center

brandunder is evidence-based. It reads public pages, shows where every conclusion comes from and never fabricates information.

Data handling

brandunder reads public pages of the website you enter, the way a visitor would. It respects robots.txt, never signs in to websites and never reads pages that are not public. It also reads robots.txt, sitemaps, structured data and, in the full analysis, a public knowledge base (Wikidata) to look for independent confirmation.

Privacy

Analyses are private to whoever ran them. A result becomes visible to others only if you create a share link, and shared results are hidden from search engines. To prevent abuse we store a salted hash of your IP address, never the raw address. Personal email addresses that appear on analyzed pages are masked in evidence. Read the Privacy policy.

Security principles

  • Only public, safe network destinations are fetched; private and internal addresses are refused.
  • Content from analyzed websites is treated as untrusted data, never as instructions.
  • Payment and account secrets live only on the server.
  • Rate limits protect both brandunder and the websites it reads.

Reporting a vulnerability

If you find a security issue in brandunder, please report it privately by email with the steps to reproduce it. Test only with analyses you created yourself, never access or change other people's data, and don't degrade the service for others. We investigate every report and fix confirmed issues. There is no paid bug bounty.

Report a vulnerability hello@brandunder.com

Analysis methodology

Each observation becomes a signal; signals become metrics; metrics become the three dimensions of the brandunder score. The scoring model is versioned, so results stay comparable over time. See How it works.

Evidence handling

Every finding is tied to the exact words, code or source it came from, and labelled as observed, inferred, outside source or unknown. A brand is never penalized for not stating something explicitly when several signals show it, and never credited for something no evidence supports.

AI limitations

Where AI is used, it only sees evidence brandunder has already collected, must cite that evidence, and its conclusions are always marked as inferred. brandunder reports AI readiness, not AI visibility: whether an assistant mentions a brand can only be measured by testing that assistant, so it is never claimed.

Data retention

Analyses are kept so you can return to them, share them and compare them after changes. You can ask us to delete your account and the analyses linked to it at any time.

Payment security

Payments are processed by Stripe through Stripe Checkout. Card details go directly to Stripe and never reach brandunder. A payment is confirmed only when Stripe confirms it, and each payment event is applied once.

Secure payments byStripe

  • Visa
  • Mastercard
  • American Express
  • Apple Pay

Responsible analysis

brandunder describes how a website communicates; it does not make statements of fact about a company, and it does not estimate revenue, traffic, rankings or conversion rates. When the evidence is not enough, the result says unknown.