Key facts
- Near AI said an attacker used a remote code execution flaw in SGLang to run commands on one inference server and reach other infrastructure.
- Near AI said names and email addresses for approximately 100 users were downloaded, plus subscription plan and account balance data for some.
- Near AI said 18 agents on an older deployment were accessed, exposing chat histories, settings and logs.
- Near AI said API keys for two Near AI Cloud customers were exposed and that it revoked both.
- Near AI said it found no indication stored Private Chat conversations or uploaded files were accessed.
- Near AI said it reported the SGLang vulnerability to the project's developers and deployed a fix across all inference servers.
Near AI disclosed a breach in which an attacker accessed account information belonging to about 100 users and data from 18 AI agents, according to a company report dated November 9, 2026. The company said the intruder used a remote code execution vulnerability in SGLang, an open-source engine that runs AI models and generates responses to user requests, to execute commands on one inference server and then reach other parts of Near AI's infrastructure.
Investigators traced the intrusion to an internal system connected to Private Chat, a feature the company had already retired. Near AI said the attacker downloaded names and email addresses for approximately 100 users, along with subscription plan and account balance information for some of them. A further 18 agents were accessed on an older deployment of Near AI Agents, where the attacker retrieved chat histories, settings and logs. Those agents were not running inside Near AI's confidential computing environment, which is designed to shield information from the company operating the servers.
Logs also exposed Near AI Cloud API keys belonging to two customers, which function as digital credentials applications use to access cloud services. Near AI said it revoked both keys but did not say whether the attacker used them. The company said it found no indication that stored Private Chat conversations or uploaded files had been accessed, while exposure of live confidential requests processed on November 8 remains a possibility identified by the investigation.