Lovable Apps Expose Sensitive User Data
A security investigation involving Lovable applications has found that tens of thousands of apps built with the platform exposed sensitive user and business data through poorly configured Supabase database integrations. Using data from the Common Crawl web archive, researchers analyzed 159,041 public Lovable projects between December 2025 and September 2026. Of those, 34,545 used Supabase, and 24,664 sent database records immediately on page load without requiring login or user interaction. While public data is not inherently unsafe, researchers were able to retrieve additional database content through read-only requests. They identified email addresses in 10,560 tables across 7,403 of the exposed sites, representing about 30 percent of the sites that loaded database records at startup. The exposed information included several high-risk datasets. A German lifestyle forum revealed more than 80,000 user profiles with full names, email addresses, dates of birth, and gender information, along with more than 200,000 chat messages. Event organizers and restaurants exposed tens of thousands of ticket and food sales records, including emails, phone numbers, postal addresses, payment information, and ticket codes. A property management company disclosed AI-generated summaries of internal emails containing sensitive details about tenants, property owners, and business partners. In one severe case, a service provider made more than 1,000 identity document records accessible, including passport numbers, expiration dates, and nationalities, creating identity theft risks. Other exposed material included contracts, invoices, personal photo albums, internal WhatsApp communications, newsletter subscriptions, waiting lists, support inquiries, and in one case 280,000 customer emails. Researchers said such exposure could facilitate phishing, ticket fraud, or identity theft, though misuse could not be confirmed. The report warned that recurring database integration patterns in Lovable applications could make it easier for attackers to replicate misconfiguration checks across many projects. It emphasized that no-code or AI-assisted development does not remove the need for expert security review, particularly before apps handle confidential data. It recommended reviewing access permissions, interfaces, and account separation before launch and after major changes, and obtaining qualified support if in-house expertise is lacking. Lovable was contacted about the findings.
