Osegi
Security·1 min read

NEAR AI says SGLang flaw exposed user account details

NEAR AI disclosed a security breach tied to an SGLang vulnerability that exposed user account details, but says it found no sign that chats or files were accessed.

Padlock icon over a digital network background symbolizing a security breach at an AI chat service.
Image: The Block
SUMODialoediviewergiftdetect

NEAR AI has disclosed a security breach in which a vulnerability in SGLang exposed user account details. The company says the incident affected account information, but it has found no indication that stored Private Chat conversations or uploaded files were accessed.

What the company says

According to NEAR AI, its investigation so far has turned up no evidence that the content users store in Private Chat was reached. The same applies to files users uploaded to the service. The distinction matters because the product is positioned around private conversations with AI models, so the question of whether chat content was exposed is central for users.

SGLang is the software component at the center of the incident. NEAR AI attributes the exposure to a vulnerability in it, rather than to a problem with the stored conversations themselves.

Why it matters

Account details can be misused even when message content stays untouched, for example in phishing attempts aimed at affected users. For a service built around privacy, any exposure of user data tests trust, and the company's statement that conversations and files appear unaffected is an early finding, not a final conclusion.

The material available does not specify which account details were exposed, how many users were affected or when the issue was fixed. Users of NEAR AI should watch for direct communication from the company about the scope of the breach and any recommended steps.

Apps worth a look

SUMO: the number puzzle

Tap the right tiles, add them up, hit the target. Easy to learn, hard to master. Free, for all ages.

Tags:near aisglangdata breachvulnerabilityprivate chat

More from this day

See all →

More in: Security

See all →