# Near AI Breach Exposes User Details and Data From 18 AI Agents

- URL: https://talkfuse.com/story/near-ai-breach-exposes-user-details-and-data-from-18-ai-agents/
- Published: 2026-11-10T20:16:03+00:00
- Updated: 2026-11-10T21:37:22+00:00
- Original publisher: Decrypt
- Original article: https://decrypt.co/382965/near-ai-breach-exposes-user-details-data-ai-agents
- Topics: AI, Cybersecurity, Technology

## Key points

- About 100 users’ account details were exposed.
- Data from 18 older AI agents was accessed.
- Two customer API keys were revoked; live request exposure remains possible.

## Summary

Near AI says an attacker exploited a software flaw in its AI model serving system, accessing account details for about 100 users and data from 18 AI agents. The breach matters because the company promotes privacy protections for AI processing, yet the flaw allowed access to systems where confidential requests were handled.

The attacker obtained names and email addresses, and some users’ subscription and balance information; the 18 agents’ chat histories, settings and logs were also accessed. Logs exposed API keys for two customers, which Near AI has revoked, though it did not say whether they were used. The company found no evidence that stored Private Chat conversations or uploaded files were accessed, but said exposure of live confidential requests remains possible. It revoked sessions and credentials, took Private Chat offline, fixed the flaw and notified affected users.

---
Summarised by TalkFuse (https://talkfuse.com/). Quote the original publisher for the reporting and link this page for the summary and discussion.