Product overview

Alibaba Cloud offers two API gateway products: Cloud-native API Gateway for general traffic and microservice scenarios, and AI Gateway for AI application scenarios. Use this guide to understand what each product does and choose the right one.

Cloud-native API Gateway

Cloud-native API Gateway unifies traffic gateways, microservice gateways, and security gateways into a single architecture, delivering 1–5x the performance of self-managed gateways. It is battle-tested at scale, proven during the Double 11 Shopping Festival, and built on years of technical expertise.

It covers core API infrastructure needs:

Service exposure and traffic control: Publish services externally, manage traffic ingress, and apply fine-grained routing policies.
Full lifecycle management of APIs: Design, deploy, version, and retire APIs from a single control plane.
Layered security: Enforce mutual TLS (mTLS), authentication, Web Application Firewall (WAF), and custom security plug-ins at the gateway layer.
Extensibility: Extend behavior with a broad plug-in ecosystem that supports hot reload — no gateway restarts required.

For more information, see Cloud-native API Gateway.

AI Gateway

AI Gateway is a unified proxy purpose-built for AI applications. It sits in front of large language model (LLM) APIs and MCP Servers, giving you centralized control over how your applications call AI services.

Key capabilities include:

Unified model access: Route requests to multiple LLM providers through a single endpoint with model switching support.
Cost and performance controls: Apply token-based rate limiting and semantic caching to reduce latency and spend.
Safety and compliance: Enforce content moderation, authentication, and web search guardrails across all AI traffic.
Observability: Monitor LLM usage, latency, and errors in one place.

For more information, see AI Gateway.

Which product should I use?

Scenario	Recommended product
Expose microservices or backend APIs to external clients	Cloud-native API Gateway
Manage traffic routing, rate limiting, and security for web applications	Cloud-native API Gateway
Proxy requests to LLM APIs (OpenAI, Qwen, etc.) or MCP Servers	AI Gateway
Reduce LLM costs with semantic caching and token-based rate limiting	AI Gateway
Enforce content moderation and safety policies across AI traffic	AI Gateway

If your application combines both general API traffic and AI service calls, you can deploy both products in the same environment — they are designed to complement each other.