NEAR AI disclosed that an attacker exploited a flaw in SGLang to enter its systems and download account details for about 100 users. The attacker also retrieved chat histories, settings and logs from 18 agents on an older deployment outside NEAR AI’s confidential computing environment. API keys for two Cloud customers were exposed and have been revoked; NEAR AI did not say whether the attacker used them.
NEAR AI said requests processed in confidential mode on November 8 may have been visible, but it found no indication that stored Private Chat conversations or uploaded files were accessed. It said it had fixed the flaw across its inference servers, rotated credentials and notified users whose account details or agents were confirmed to have been accessed.
