Higress is a cloud-native AI gateway based on Istio and Envoy, offering extensibility with Wasm plugins and a user-friendly console. It provides high availability and load balancing capabilities for various services.
Higress is a cloud-native API gateway based on Istio and Envoy, extended with Wasm plugins (Go/Rust/JS). It offers ready-to-use plugins and a console. Born within Alibaba, it addresses Tengine reload issues and load balancing for gRPC/Dubbo. Alibaba Cloud's API gateway product is built on Higress, ensuring high availability.
Higress's AI gateway supports mainstream model providers and self-built models (vllm/ollama). It supports AI businesses like Tongyi Qianwen APP, Bailian API, and the PAI platform, serving AIGC enterprises and AI products.
Key features: AI Gateway capabilities (observability, load balancing, rate limiting, caching), MCP Server hosting for AI Agents, Kubernetes ingress controller (Gateway API support coming soon), microservice gateway (Dubbo, Nacos, Sentinel integration), and security gateway (WAF, authentication strategies).
Core advantages: Production-grade, streaming processing, easy extensibility via Wasm, and secure/easy to use with a UI console and automatic certificate issuance.
alibaba/higress
October 27, 2022
March 28, 2025
Go