Arthur Engine releases the world's first open source real-time AI evaluation engine to ensure AI security for free

Arthur recently launched Arthur Engine, the world's first open source real-time AI evaluation engine, aiming to help development teams monitor, debug and optimize generative AI and traditional machine learning (ML) models. Unlike traditional monitoring tools, Arthur Engine can run locally to ensure data privacy and security, while completely free, providing efficient and real-time evaluation capabilities for AI systems.

The engine has real-time detection, active protection and highly customizable monitoring capabilities, and supports mainstream models including GPT, Claude, Gemini, and other open weights and traditional ML models.

The core features of Arthur Engine include:

• Real-time AI evaluation, instant detection of failures and prevent model degradation;

• Active protective measures, automatic intervention when abnormalities are detected;

• Customized evaluation indicators are suitable for different application scenarios.

Cherie Xu, technical director of Arthur, said that he hopes that through open source Arthur Engine, developers can more easily build secure, transparent and compliant AI systems and promote the healthy development of AI technology. At present, the engine is open on GitHub, and users can also join the Arthur platform's waiting list to get more information.