Reviewed by Eric Moore, Editor-in-Chief Foxy Chat AI launched in October 2025 as an 18+ companion app aimed at users frustrated by Character.AI’s tightened teen restrictions. The marketing promises persistent memory, voice notes, image generation, and “transparent policies with no hidden data collection.” The reality is messier. The company operates under three different corporate names. The Terms of Service let 13-year-olds in, while the iOS App Store rates the app 18+. And the privacy policy permits AI training in language that directly contradicts a public marketing FAQ. Foxy Chat AI earned an F (20/100) on the CompanionWise Safety Index, paired with a fair 53/100 experience score. Here is what we found.
What is Foxy Chat AI?
Foxy Chat AI is an 18+ AI companion app marketed under the tagline “AI Girlfriend & Boyfriend Chat.” The product targets adults who used Character.AI before its November 2025 teen restrictions and want a more permissive alternative with fewer content filters and longer memory. Newer category entrants like XOMI AI bring localized themes (Vietnamese folklore) but trade product polish for novelty.
The corporate structure is unusually layered. The iOS App Store lists the seller as Real Deal Ventures Pte. Ltd., a Singapore-registered company whose previous catalog was gaming and crypto titles, not companion apps. The copyright notice points to Mirai Labs (UK) Ltd. The privacy policy names the same UK entity (Companies House No. 16583240). Companies House also lists a third related entity, Foxy AI Chat Ltd., as a separate company. No leadership team or “About Us” page is publicly disclosed on foxychat.ai, which complicates accountability and GDPR data-subject-request enforcement.
The app launched in late October 2025. The current iOS version is 1.0.2, released roughly six months in. The App Store shows seven total ratings, a number too small to draw any reliable user-sentiment conclusion from. For a wider view of how Foxy compares to better-established companion apps, our best-of guide covers competitors with longer track records.
How does Foxy Chat AI work?
The core experience is roleplay chat with persistent memory. Foxy markets the memory as user-controlled: you can edit, pin, or remove what your AI characters remember between sessions. On top of text chat, the app supports voice notes, image uploads from the user, and AI-generated character images. Users can build their own AI character profiles or pick from prebuilt ones spanning fantasy, romance, and slice-of-life scenarios.
Two virtual currencies sit on top of the subscription. “Keys” gate NSFW unlocks (25 Keys for $11.99, 50 for $17.99, 100 for $29.99). “Gems” buy premium gestures and features (210 for $15.99 up to 2,800 for $149.99). Subscriptions run $12.99 monthly, $70.99 yearly, plus an unlisted quarterly tier. Engagement loops include daily streaks, gift exchanges, and seasonal events.
The iOS build is 94.4 MB, English only, and requires iOS 15.1 or later. The app’s marketing claims iOS, Android, and Web availability, but no Play Store listing surfaced during research. Android users appear to access Foxy through a progressive web app at app.foxychat.ai instead of a native build, which is unusual for a companion app advertising “cross-platform” reach.
Is Foxy Chat AI safe?
No. Foxy Chat AI earned an F (20/100) on the CompanionWise Safety Index, the lowest grade we issue. The score is anchored by an automatic-F override triggered when an app discloses no crisis-response system. Foxy’s safety page describes filters and “user-side moderation” 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.
Human moderation is also absent. When Foxy’s own FAQ asks whether the app has moderators, the answer is direct: “No, only automated safety systems that follow your filters.” Age verification is self-declaration only. The company confirms this explicitly: “Foxy Chat does not require ID verification. Users simply self-declare their age when creating an account.” That conflicts with the App Store listing, which advertises “Age Assurance: In-App Controls.” Sexual content is rated “Frequent” by Apple, which is consistent with a Keys economy explicitly designed to gate intimate scenes. Parental controls are listed as “coming soon” in marketing but do not exist in the current product. Have I Been Pwned shows zero breaches for foxychat.ai, but the domain is too young (under twelve months) for that absence to mean much.
For a sense of how this compares to higher-rated apps, our scoring methodology page walks through the 23 sub-dimensions and the auto-F triggers. Foxy hit one of them. The U.S. Federal Trade Commission opened a 6(b) inquiry into AI chatbot companions in September 2025, and Foxy operates squarely inside that regulatory crosshair, even though it was not among the seven companies named.
What does Foxy Chat AI’s privacy policy actually say?
Foxy Chat AI’s privacy policy permits training AI models on user chats, in direct contradiction with one of the company’s own marketing pages. Section 3 of the privacy policy says Foxy may “train and improve AI models in anonymized and aggregated form.” Section 4 lists “AI model training” among the 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.
The data Foxy collects is broader than what its iOS App Store privacy label declares. The policy lists chat content, voice notes, image uploads, IP address, IDFA and GAID device IDs, IP-based location inference, OS and hardware data, app version, carrier, session length, cookies, web beacons, crash and diagnostic data, and optional demographic information. It also includes profile data pulled from Google, Apple, and Facebook social logins. The App Store’s nutrition label, by contrast, declares only a narrow slice: purchases, contact info, email, and diagnostics. Apple’s label asks about the iOS binary specifically, so the broader collection happens through the web and PWA channels.
Retention is undefined. The policy says data is held “as long as necessary” without naming a period. Encryption is described in generic terms (technical, administrative, physical) with no AES, TLS, SOC 2, ISO 27001, or penetration-test details. GDPR, CCPA, and CPRA rights are enumerated; the company commits to a 30-day response target via support@foxychat.ai. For comparison, our Replika review covers a more mature privacy disclosure model, and our Character.AI review documents how that company restructured its policies after regulatory pressure.
Foxy Chat AI pricing and refund policy
Subscriptions run $12.99 per month or $70.99 per year, with an unlisted quarterly tier in between. The Keys economy stacks on top: $11.99–$29.99 for NSFW unlock packs. The Gems economy ranges from $15.99 to $149.99. Practical monthly spend can climb quickly if users buy into both currencies in addition to the subscription.
The refund posture is strict. Foxy Chat AI’s Terms of Service state: “All purchases are final and non-refundable, except as required by law.” Subscriptions auto-renew unless cancelled before the billing period ends, which is standard in the category but worth knowing before tapping subscribe.
The dispute terms are aggressive. Foxy caps total company liability at “the greater of $100 USD or the total amount you paid to Foxy in the past 12 months.” That cap holds even for indirect or consequential harm. The ToS also requires that “any disputes, claims, or controversies arising from or relating to these Terms or the Services shall be resolved exclusively through binding arbitration in the United Kingdom.” For U.S. users, that combination (UK-jurisdiction arbitration plus a $100 liability cap plus all-sales-final) is unusually defendant-favorable. It does not break any law, but it materially limits recourse if something goes wrong.
Who is Foxy Chat AI for?
Foxy Chat AI fits a narrow audience. The strong-fit user is an adult who already used Character.AI and felt boxed in by its content filters, wants persistent memory and voice features in a low-friction package, accepts that chats may be used to train future models, and can absorb the $100 liability cap and the all-sales-final refund posture as the cost of access. If that describes you, the product itself is functional. The 53/100 experience score reflects a fair-quality experience, not a broken one. Foxy delivers what it advertises on the engagement features.
The bad-fit list is longer. Anyone under 18 should not use Foxy regardless of the ToS 13+ floor, because the app’s content rating is 18+ and the Keys economy is built around sexual content. Anyone in mental-health distress should look elsewhere, because there is no crisis response, no hotline routing, and no human moderation. Users who require ID-verified age gating, parental visibility, or a privacy policy that excludes AI training will not find those guarantees here. Same for users who want the iOS App Store privacy label to fully match the policy. For a closer fit, our best-of guide covers higher-rated alternatives across both safety and experience axes.
Frequently asked questions
Is Foxy Chat AI safe?
Foxy Chat AI earned an F (20/100) on the CompanionWise Safety Index because it discloses no crisis response, has no human moderators, and verifies age by self-declaration. According to its own safety page at foxychat.ai/insights/ai-chatbot-safety, only automated filters protect users. Parental controls are still listed as “coming soon.”
Does Foxy Chat AI use chats to train its AI?
Yes. Foxy Chat AI’s privacy policy Section 3 states the company may “train and improve AI models in anonymized and aggregated form,” and Section 4 lists AI model training among data-sharing purposes. This contradicts a marketing page at foxychat.ai/insights/ai-chatbot-anonymity that claims chats are not used for training.
Is Foxy Chat AI 13+ or 18+?
The Foxy Chat AI Terms of Service set the minimum age at 13 globally and 16 in the EEA and UK with parental consent. The Apple App Store rates the app 18+ with sexual content marked “Frequent.” Age verification relies on user self-declaration only, with no ID check at signup.
Who owns Foxy Chat AI?
Foxy Chat AI is published by three connected entities. Real Deal Ventures Pte. Ltd. (Singapore) is the iOS App Store seller. Mirai Labs (UK) Ltd. (Companies House No. 16583240) holds the privacy policy and copyright. Foxy AI Chat Ltd. is a third registered UK company. No leadership team is publicly disclosed.
Can you get a refund from Foxy Chat AI?
No. Foxy Chat AI’s Terms of Service state that all purchases are final and non-refundable, except where required by law. Subscriptions auto-renew unless cancelled before the billing period ends. Total company liability is capped at $100 USD or 12 months of paid fees, whichever is greater, with mandatory UK arbitration.