Braintrust

Profile narrative

Braintrust operates as an AI platform that enables customers to deploy and manage machine learning models via programmable interfaces. The service is also referred to as Braintrust AI platform in public communications. Customers authenticate to the platform using API keys that are stored within the provider’s infrastructure. By handling these credentials, Braintrust facilitates secure access to its AI services for a range of users. The platform’s design centers on providing scalable AI capabilities while managing credential lifecycle for its clientele.

On May 4, 2026, Braintrust identified unauthorized access to an Amazon Web Services account that housed customer API keys. Upon detection, the company immediately locked down the compromised AWS account to prevent further intrusion. Internal credentials associated with the account were rotated, and affected users were notified of the potential exposure. The breach resulted in one confirmed instance of customer impact and prompted an investigation into three additional customers exhibiting unusual AI usage patterns. No evidence was found indicating broader data exfiltration beyond the accessed API keys.

Investigation traced the intrusion to compromised cloud credentials, aligning with MITRE ATT&CK technique T1078.004 (Valid Accounts: Cloud Accounts). In response, Braintrust engaged external incident response specialists to assist with containment and remediation efforts. The organization subsequently implemented additional safeguards aimed at strengthening cloud credential management and monitoring. These actions reflect a focus on rapid containment and proactive improvement of security posture following the incident. Braintrust continues to operate its AI platform while maintaining heightened vigilance over credential security.

Incidents

1 incident linked to this organisation.