This project is a high-performance, distributed API gateway designed to manage, secure, and observe traffic for microservices, serverless functions, and artificial intelligence model providers. It functions as a dynamic service proxy and cloud-native ingress controller, centralizing policy enforcement and traffic routing through a unified configuration interface that synchronizes state across multiple nodes in real time.
The platform distinguishes itself through a highly extensible architecture that utilizes a high-performance scripting engine to execute modular logic directly within the request lifecycle. It provides specialized capabilities for modern AI workflows, including model request proxying, token-based budget enforcement, content moderation, and agentic workflow tracing. Furthermore, it supports complex multi-protocol environments by bridging diverse communication standards, including gRPC and various binary protocols, without requiring additional sidecar processes.
Beyond its core proxying functions, the gateway offers a comprehensive suite of traffic management and security tools. It handles authentication and authorization through multiple strategies, including token validation and identity provider integration, while maintaining granular control over TLS policies and secret management. The system also provides robust observability through distributed tracing, metrics exporting, and detailed request logging, ensuring visibility into both standard API traffic and complex AI-driven interactions.
The software is designed for containerized environments and can be deployed using standard container images, with full support for translating Kubernetes ingress resources into live routing rules.