NEAR AI Says SGLang Flaw Exposed User Account Details

NEAR AI has disclosed a security breach that exposed user account details. It traces the problem to a vulnerability in SGLang. The content users care about most, the company says, looks untouched.

What NEAR AI says happened

The company says the SGLang vulnerability let account details leak. That's the confirmed part, and it's about all we've got. The available facts don't say how many users were affected, which specific fields were exposed, or when the breach began.

On the more sensitive side, NEAR AI says it found no indication that stored Private Chat conversations were accessed. Same goes for uploaded files. Read that wording closely. "No indication" isn't "confirmed safe." It means the investigation hasn't turned up evidence so far. Not the same claim.

What we don't know yet

• How many accounts were exposed
• Exactly which account details leaked
• How long the vulnerability was open
• Whether the SGLang flaw has been patched

Why it matters

If you use NEAR AI, watch for phishing. Exposed account details are exactly what scammers use to build convincing messages, so treat unexpected emails or links with suspicion and review your account security settings. The bigger point reaches past one company. AI platforms lean on open inference tooling like SGLang, and a bug in that layer can ripple into the products built on top of it. For a service that sells privacy as a feature, the real test is how fast it publishes details, and whether its "no access" claim holds up once someone looks harder.