Israeli security research firm Red Access says it found thousands of apps built with AI coding tools leaking sensitive customer data, in the latest sign that the vibe-coding boom is racing ahead of basic application security. The firm scanned 380,000 publicly accessible apps built on Lovable, Replit, Base44 and Netlify, and reported that roughly 5,000 of them shipped with no authentication at all.
Of that group, about 2,000 apps were actively exposing data to anyone who stumbled onto the URL. "So about 40% of those or about 2000 apps were exposing sensitive data to anyone who happened to find the URL," Jaeden Schafer said on the AI Chat Daily podcast, walking through the findings.
The categories of data Red Access surfaced read like a worst-case compliance audit. Researchers found conversations between doctors and patients at a long-term care facility for children, a school app containing lesson recordings and student records, a vacation planning tool with customer details, customer chatbot logs and internal corporate strategy decks. In one moment of irony, they also pulled up a security company's own incident response data sitting unprotected on the open web.
Key facts
- 01Israeli security firm Red Access scanned 380,000 publicly accessible apps built on Lovable, Replit, Base44 and Netlify and found about 5,000 with no authentication.
- 02Roughly 2,000 of those apps were actively exposing sensitive data to anyone with the URL, including doctor-patient conversations at a long-term care facility for children.
- 03Exposed material also included a security company's incident response data, school lesson recordings and student records, customer chatbot logs and internal corporate strategy decks.
- 04Base44 has retroactively made all projects created before a certain date private by default, while Replit is pushing back on Red Access for giving it only 24 hours before going public.
The vendors are reacting unevenly. Replit is pushing back on the disclosure timeline, with its chief executive arguing that Red Access only gave the company 24 hours before taking the findings to the press. Schafer said the complaint has some merit, noting Replit's position is essentially "why don't you guys give us more time to like fix it before you went live with it."
“there's a huge difference between AI building you a working app and AI building you a secure app that isn't going to leak your data.”— Jaeden Schafer
Base44 appears to have moved more aggressively. Schafer, who holds a Base44 account, said he received notice that older projects had been quietly locked down. "all of the all projects on base 44 made before a certain date were automatically made private If you want to go make it public again, you can go do it, but you got to do like some security stuff," he said, describing the platform's response.
The pattern of legacy projects carrying the worst exposure is showing up across the category. Lovable hit a similar wall earlier, where projects created before November 2025 carried security gaps that newer builds did not. The fix industry-wide is converging on the same answer: flip older user-generated apps back to private by default, and require the owner to clear a security checklist before re-exposing them to the public internet.
The episode highlights the gap between what AI coding agents are good at today and what production software actually requires. "there's a huge difference between AI building you a working app and AI building you a secure app that isn't going to leak your data," Schafer said, framing authentication and access control as the next problem the vibe-coding platforms have to solve before regulated data ever touches their tools.
For now the practical implication for builders is straightforward. Anyone who shipped a Lovable, Replit or Base44 project earlier this year — particularly one handling health, education or internal corporate information — should assume it was scanned, check whether the host platform has already forced it private, and treat re-publication as a security exercise rather than a one-click toggle.
Working on something we should cover, or seeing a story we missed? Send leads, documents, or feedback to hello@aichatdaily.com. For sensitive tips, see our secure tips page for Signal and PGP options.
Spotted an error? Email hello@aichatdaily.com with the URL and the issue, or read our full corrections policy.
