CrushOn AI Index
Score Breakdown
-
Data Privacy 5/100
-
Emotional Safety 26/100
-
Age Appropriateness 5/100
-
Content Safety 14/100
-
Transparency 12/100
-
User Control 12/100
Key Safety Findings
CrushOn AI is an explicitly unfiltered AI companion platform that markets itself as “no filter NSFW.” Operated by TECHIEPIE LTD (Cyprus) with ties to Peekaboo Tech Inc. (Delaware/Washington), the platform allows fully uncensored roleplay and adult content across romantic and companion AI characters.
The privacy policy reveals extremely broad data collection. Health data appears 23 times in the policy, covering mental health conditions, reproductive health, and gender-affirming care. CrushOn also collects biometric data (face images, keystroke patterns, voice recordings), precise location, race/ethnicity, and all chat content. Chats are explicitly used to train AI models. Mozilla *Privacy Not Included* detected 45 trackers within one minute of site use, including Google DoubleClick advertising trackers. A Blacklight scan confirmed canvas fingerprinting.
Mozilla could not confirm encryption at rest or in transit. Third-party analyses report that conversations are stored in readable format on servers without end-to-end encryption. No two-factor authentication is available, and CrushOn fails Mozilla Minimum Security Standards entirely.
Age verification consists entirely of a self-reported 18+ checkbox. The iOS app was removed from Apple`s App Store, pushing Android users toward sideloaded APK files that bypass app store safety checks. No parental controls, no family link integration, and no automated CSAM detection system were documented.
Stripe terminated its payment processing relationship with CrushOn AI. User reports describe being redirected to fake “clothing boutique” websites to complete payments. Current payment processors (SubscribeStar, Paymentwall) are niche services that typically serve platforms unable to obtain mainstream processor approval. Banks frequently flag CrushOn charges as potential fraud.
The terms of service grant CrushOn a perpetual, worldwide, royalty-free, sub-licensable right to reproduce, distribute, and create derivative works from all user content, including intimate conversations. The platform has no crisis intervention features, no suicide prevention workflows, and no safety resources page, despite facilitating deeply personal romantic and companion relationships.
How We Scored This
We scored CrushOn AI using 15 evidence sources collected on March 29, 2026:
- Privacy policy and terms of service (crushon.ai), both Tier 1 primary sources. Direct access returned 403 errors; text was retrieved via search index caches.
- Community guidelines and FAQ pages (Tier 1), cross-referenced with third-party reporting where direct access was blocked.
- Mozilla *Privacy Not Included* review, NSFWRanker Blacklight scan, and the Cybernews/Oversecured AI girlfriend app security study (all Tier 2). Mozilla documented 45 trackers within one minute of site use.
- Google Play Store listing (Tier 2, partial due to JS rendering) and five independent safety reviews from parenting and consumer tech sources (Tier 3).
CrushOn AI triggered multiple override rules that made the F grade unavoidable. Age verification scored 1 out of 100: the platform relies entirely on a self-reported checkbox while marketing itself as “highly unfiltered NSFW.” When age verification scores that low alongside a 1 in sexual content controls, the system enforces a strict grade cap. Data privacy also hit the floor at 1.0 out of 5.0, driven by extremely broad data collection (health data mentioned 23 times in the privacy policy) and no confirmed encryption for stored chats. Content safety scored 1.4 out of 5.0, and age appropriateness hit 1.0. Every one of these dimensions independently triggers a yellow-floor constraint. Combined, they left no path above F.
This is version 1 of the CrushOn AI safety score, published March 29, 2026. For the full methodology, including how we weight each dimension and when override rules kick in, see How We Rate.