Over 100 Fake AI Sites: A $249 Toolkit Lowers the Barrier for Subscription Fraud

Malwarebytes Threat Intelligence has identified more than 100 fake AI subscription sites that share the same commercial infrastructure — a $249 one-time purchase — selling paid plans up to over $2,000 a year. The network impersonates real products such as GPT-6 Astra, DaVinci Resolve, Omegle, PixAI, and OpenCut, alongside invented brands. The key mechanism is the use of genuine Google sign-in screens that generate apparent legitimacy while hiding operators traceable only through free Gmail addresses with similar naming patterns.

How the Network Works: Commercial Toolkit and Mass Impersonation

Malwarebytes' discovery centers on a recurring pattern: more than 100 AI subscription sites sharing identical files, identical checkout pages, identical account settings, and developer email addresses with similar naming patterns. The source states explicitly: "We found more than 100 subscription websites linked through the same toolkit and closely related developer details." This technical convergence points to a single operator or a tightly connected group, not a multiplicity of independent actors.

The underlying infrastructure is a legitimate commercial toolkit, not designed for fraud but repurposed in that context. According to Malwarebytes, the kit costs $249 as a one-time purchase, with additional templates at roughly $2 each. The vendor advertises that customers can "launch a product in an hour." This low barrier to entry turns the creation of seemingly professional subscription sites into an operation within reach of minimal budgets.

The analyzed sites fall into two categories: those impersonating existing products and those promoting entirely invented brands with unverifiable operator information. Among the impersonated brands are GPT-6 Astra, DaVinci Resolve, PixAI, OpenCut, and Omegle — the latter a service shut down in 2023, making its impersonation particularly emblematic of the superficial credibility crafted for the scheme.

The Price of Fraud: From $89 a Month to Over $2,000 a Year

The identified pricing tiers cover a wide spectrum. The fake GPT-6 Astra site offers three levels: $89, $169, or $249 a month. Other sites in the network propose subscriptions from under $10 a month up to annual plans exceeding $2,000. One reported case claimed over 12 million users, a claim lacking independent evidence.

The pricing structure mimics that of legitimate AI services: differentiated tiers, recurring billing, emphasis on perceived value. The substantial difference lies in the absence of any way to test the product before payment. Landing page controls do not work without authentication, and sign-in is the mandatory first step toward purchase.

The Google Sign-In Trick: Authentic but Misleading

The mechanism of greatest technical interest is the integration of real Google OAuth. Users see a genuine Google URL and enter their password on Google's servers, not on a spoofed page. Malwarebytes clarifies the point: "A genuine Google sign-in page does not confirm that the service is official or connected to the brand it displays. It only confirms that Google is handling the login and passing the approved information to an outside application."

The sites request limited permissions — name, email, and profile photo — but not access to Gmail or Google Drive. This apparently prudent choice reinforces the illusion of legitimacy: the user perceives a "safe" service because it does not request invasive permissions, without understanding that the basic data granted still identifies an unverifiable operator.

Developer information accessible after authorization shows free Gmail addresses with similar naming patterns across the different sites, not corporate addresses of the impersonated brands. Malwarebytes observes: "A free email address is not proof of wrongdoing, but it should raise questions when a service claims to have millions of users or presents itself as an established company."

"We found more than 100 subscription websites linked through the same toolkit and closely related developer details." — Malwarebytes Threat Intelligence

Traces Left by the Toolkit: Demo Material Still Visible

A relevant aspect for investigative reading is the presence of original toolkit demo material still in active sites. Many contain generic testimonials, default menu items, and placeholder content not replaced. One site labeled a demo list of companies as its own clients; on another the word "boilerplate" remained in the name of a paid plan.

These details reveal not only haste or carelessness by the operators, but also how the toolkit's automation reduces the need for specific technical skills. Superficial customization — logo, product name, colors — is enough to generate a commercial facade, while the operational substance remains identical across the entire network.

Verification of corporate ownership proved impossible: no registered company name, business address, or independent ownership detail was found. Support contact addresses are free webmail accounts, not matching the domains of the impersonated brands.

What to Do Now

For users encountering unknown AI subscription sites, the Malwarebytes brief suggests specific checks. Verify that the support contact email matches the corporate domain of the claimed brand: the presence of free Gmail accounts instead of professional addresses is a documented signal of non-official status. Check whether the product is testable without mandatory authentication: the inability to access features before sign-in is a recurring pattern in the examined network.

Examine user-count claims on a case-by-case basis: a site declaring millions of users should present independently verifiable evidence, absent in the analyzed cases. Look for unreplaced demo content, such as generic testimonials or default menu items, indicating use of uncustomized templates. Finally, confirm that the impersonated brand actually exists: Omegle, cited among the fake sites, is a service shut down since 2023.

For payments, the brief does not document specific refund methods or chargeback procedures. The practical recommendation is to halt any transaction when one or more of these indicators are present, and report the site through the abuse channels of the domain registrar or the payment platform used.

Why This Matters

The dossier does not specify the total number of victims nor the overall economic scale of the fraud. It does not emerge whether the promised AI services are delivered in any form or are purely nonexistent. The brief does not document specific remedial measures nor ongoing legal actions or official investigations.

The specific name of the toolkit vendor is not mentioned in the examined material. The dossier also does not specify whether user-uploaded files are used for further purposes. The exact identity of the operator or group remains undetermined.

What the report documents clearly is evidence of operational scalability: a $249 investment plus roughly $2 per additional template allows the generation of more than 100 sites with a professional appearance, SSL, integrated billing, and Google authentication. This model challenges traditional trust indicators — secure certificates, star reviews, integration with recognized platforms — that users have been conditioned to consider sufficient.

Frequently Asked Questions

Do the sites steal Google passwords?
No. Sign-in occurs on genuine Google URLs; the password does not pass through the fake sites' servers. The risk is the sharing of name, email, and profile photo with unverifiable operators, followed by payments for potentially nonexistent services.

Is the toolkit illegal?
The toolkit is a legitimate commercial product. The illicit use is by those who redeploy it with false branding and fraudulent intent, not by the original vendor according to the documented brief.

How to recognize these sites?
Documented anomalous elements include: free Gmail addresses instead of corporate brand addresses; inability to test the product before sign-in; claims of millions of users without verifiable evidence; presence of unreplaced demo content.

Information is based on the cited source and current as of publication.

Sources

• https://www.malwarebytes.com/blog/threat-intel/2026/09/the-fake-sites-using-a-cheap-toolkit-to-sell-2000-ai-subscriptions