Evidence, not badges.
A testing company that overstates its own results is useless. This is the public promise that keeps our marketing honest, including this website.
- “Compliant” (WCAG, ADA, EAA or anything else)
- “Certified” or “agent-ready certified”
- “Agent-ready guaranteed” or “bug-free”
- “X% coverage” or any recall figure
- “Simulated blind / elderly users”
- “Trusted by N teams” without a named, consenting source
- What we observed, on which date, with which model versions
- Precision per persona as confirmed ÷ flagged, with n and 95% CI
- The fixed limitations statement on every report
- That consumer-agent classes are behavioural approximations
- That narration and video are AI-generated (EU AI Act Art. 50)
- Sources and dates for every third-party statistic
Why this matters
The US Federal Trade Commission requires competent and reliable evidence before a claim is made. In 2025 it ordered Workado to stop advertising a “98% accurate” AI detector that measured 53% in testing, and fined accessiBe $1 million over accessibility-compliance claims and undisclosed paid reviews.
We cleaned our own house first
On 25 Sep 2026 we removed from this site every claim we could not back up: SOC 2 Type II, ISO 27001, HIPAA and EU AI Act “certified” badges, 99.99% uptime SLAs, “500+ teams”, “thousands of developers”, download counts, speed multiples, and documentation for a CLI that did not exist. Our actual security status is on the trust centre.
How it is enforced
Every public claim lives in an internal claims register with its source, sample size, confidence interval, owner and review date. A claim without evidence cannot be published. Reviews we solicit disclose any incentive.
Spotted a claim that breaks this policy? Email [email protected] and we will correct it publicly.