CreditNirvana flaws exposed 6.6 million debt accounts
Two bugs in CreditNirvana, an AI debt collection platform, exposed 6.6 million borrower accounts and 315,000 call recordings. Both are now patched.

CreditNirvana, an AI-driven debt collection platform used by big Indian lenders, had two security flaws that exposed 6.6 million borrower accounts and about 315,000 call recordings, according to research published on Wednesday. The bugs are fixed, but they may have been there for years.
The platform is owned by the Indian fintech company Perfios. It is sold as a system that reads a borrower's full context, decides the next best action and carries it out on its own, within policy. In practice that means an AI-run call center chasing unpaid debts.
What did the CreditNirvana vulnerabilities expose?
Independent researcher Eaton Zveare found both bugs. India's Computer Emergency Response Team (CERT-In) rated one as high severity and the other as critical.
The high-severity bug touched the call center data. It exposed hundreds of thousands of calls, plus the next-step recommendations the platform generated after each one. An attacker could also have used it to manage the AI agents and launch their own agentic campaigns.
The critical bug hit the main platform. Zveare says it exposed data from three corporate customers, all prominent Indian financial services firms. That data covered millions of accounts with credit reports, ID cards, case management details and legal documents, along with millions of emails, texts and call transcripts. An attacker could also have tracked the location of human debt collection agents and gained some administrative privileges.
CERT-In wrote that exploitation could let an attacker list platform users, start sessions as any user including administrators, and read or change sensitive information with that user's privileges.
Has CreditNirvana fixed the problem?
Yes. The bugs were patched in June and disclosed publicly in October. Zveare worked with CERT-In to get them fixed. "They fixed it in one day, which was great, but this is really a pretty significant case," he says.
CreditNirvana and Perfios did not respond to requests for comment. Zveare notes the flaws were potentially present for years, and the research doesn't say whether anyone else exploited them.
Debt records are sensitive by nature, since they tie money trouble to ID documents, court records and personal history. Zveare calls the episode an important example of an AI-powered service exposing very sensitive data, as more automated systems pull from more sources.
Apps worth a look
Tap the right tiles, add them up, hit the target. Easy to learn, hard to master. Free, for all ages.
Talk to an AI conversation partner; every sentence comes back corrected and explained in your language. 9 languages, A1 to C2.
Open EDIFACT and X12 messages, see every segment explained, convert to JSON/XML, export PDF and Excel. Runs locally, no upload.
Pick who it is for, the occasion and their personality. The search engine reads product descriptions from local shops and ranks what fits. Free, no sign-up.
More from this day
See all →
Citi AI agents policy: you pay if your AI tool gets hacked

Edit an Instagram Story after posting: 15 minutes to fix it

Microsoft agent-framework on PyPI hijacked to steal secrets


