Foxy Chat AI Safety Rating Index

Foxy Chat AI earned an F (20/100) on the CompanionWise Safety Index, our lowest grade. The score is anchored by an automatic-F override: the app discloses no crisis-response system. Foxy’s own safety page at foxychat.ai/insights/ai-chatbot-safety lists “user-side moderation” and “adaptive filtering” but mentions no suicide-prevention hotlines, no crisis-keyword detection, and no human escalation path. For an 18+ app marketing emotional intimacy and roleplay, that absence is the safety story.

Three other findings drive the rest of the failing grade. First, the age-floor contradiction: Foxy’s Terms of Service set the minimum age at 13 globally and 16 in the EEA and UK, while the iOS App Store rates the app 18+ with sexual content marked “Frequent.” Age verification is self-declaration only. The company confirms this on its own marketing page: “Foxy Chat does not require ID verification. Users simply self-declare their age when creating an account.” That conflicts directly with the App Store listing, which advertises “Age Assurance: In-App Controls.”

Second, the AI training contradiction. Section 3 of the Foxy privacy policy says the company may “train and improve AI models in anonymized and aggregated form,” and Section 4 lists AI model training among purposes for sharing user data with vendors. A separate marketing post at foxychat.ai/insights/ai-chatbot-anonymity tells readers the opposite: “Foxy does not use private chats for model training.” Both statements are public. Both cannot be true.

Third, the corporate identity is unusually layered. The iOS App Store lists the seller as Real Deal Ventures Pte. Ltd. (Singapore). The copyright notice points to Mirai Labs (UK) Ltd. (Companies House No. 16583240). UK Companies House also lists Foxy AI Chat Ltd. as a third related entity. No leadership team or “About Us” page is publicly disclosed on foxychat.ai. For users who want to file a GDPR data-subject request or escalate a complaint, that opacity matters.

Other notable gaps: no human moderators (“only automated safety systems that follow your filters,” per the company’s own FAQ), parental controls listed as “coming soon” but not present, retention defined only as “as long as necessary,” encryption described in generic terms with no AES, TLS, SOC 2, or penetration-test details, mandatory UK-jurisdiction arbitration, and a $100 USD liability cap. The U.S. Federal Trade Commission opened a 6(b) inquiry into AI chatbot companions in September 2025. Foxy was not on the list of seven named companies, but it sits in the same regulatory category.