Builderoom

September 28, 2026

View repo

A cloud-native API and AI gateway built on NGINX and etcd, offering dynamic routing, load balancing, and hot-reloadable plugins.

Apache-2.0
安装命令(说明)bash
curl -sL https://run.api7.ai/apisix/quickstart | sh
最小示例代码python
curl -i "http://127.0.0.1:9180/apisix/admin/routes/1" -X PUT -d '
{
  "uri": "/get",
  "upstream": {
    "type": "roundrobin",
    "nodes": {
      "httpbin.org:80": 1
    }
  }
}'

curl "http://127.0.0.1:9080/get"

Product facts

Type
API, SDK & Library
Category
Development & Engineering › API Integration
Pricing
Open source
License
Apache-2.0
Launched
Sep 28, 2026
Listed
Sep 28, 2026
Website
github.com

Overview

GitHubGitHub

Apache APISIX is a dynamic, real-time, high-performance API gateway project maintained by the Apache Software Foundation. It is written in Lua and built on top of NGINX and etcd. ## What it is APISIX serves as a traffic entry point to process business data, handling both traditional north-south traffic and east-west traffic between services. It can be deployed in environments ranging from bare metal to Kubernetes, and can also function as a Kubernetes ingress controller. ## Problems it solves In microservices architectures, traditional API gateways often require restarts for configuration updates and have limited scalability. With dynamic routing and hot-reloadable plugins, APISIX allows you to continuously update configurations and plugins without interrupting services. ## Key capabilities - **Traffic management**: Provides load balancing (round-robin and consistent hashing), dynamic upstreams, canary releases, blue-green deployments, circuit breaking, rate limiting, and more. - **Multi-protocol support**: Beyond HTTP/HTTPS, it supports proxying and transcoding for TCP/UDP, Dubbo, MQTT (3.1.x and 5.0), gRPC, gRPC Web, WebSocket, HTTP/3 (QUIC), and other protocols. - **AI gateway**: Implements AI proxying via its plugin system to route traffic uniformly to different LLM providers; supports load balancing, retries, and fallbacks across multiple LLMs; offers token-based rate limiting to control costs; and includes an `mcp-bridge` plugin that converts stdio-based MCP servers into HTTP SSE services. - **Fully dynamic architecture**: Supports dynamic certificate loading, request and response rewriting, automatic filtering of unhealthy nodes via health checks, traffic mirroring, and proportional traffic splitting. - **Security and observability**: Built-in authentication and authorization, malicious attack protection, metrics collection, monitoring and alerting, and service governance. ## Who it's for Ideal for teams managing APIs and governing traffic in cloud-native or microservices architectures, as well as scenarios requiring an API gateway as a unified entry point for AI workloads.

Key features

  • Dynamic routing and hot-reloadable plugins: Continuously update configurations and plugins without restarts, meeting API management needs in microservices architectures.
  • Multi-protocol proxying: Supports proxying and transcoding for HTTP, TCP/UDP, Dubbo, MQTT, gRPC, WebSocket, HTTP/3, and more.
  • AI gateway capabilities: Route traffic to different LLM providers through a unified interface, with cross-model load balancing, retries, and fallbacks.
  • Token-level rate limiting: Apply rate limits based on tokens to control AI invocation costs and protect upstream models.
  • MCP Bridge plugin: Seamlessly convert stdio-based MCP servers into scalable HTTP SSE services.
  • Load balancing and health checks: Supports weighted round-robin and consistent hashing load balancing, automatically filtering out unhealthy upstream nodes.
  • Traffic splitting and canary releases: Gradually direct traffic to different upstream services by proportion, supporting A/B testing and canary deployments.
  • Dynamic SSL certificates: Load and update SSL certificates dynamically without restarting.
  • Request and response rewriting: Rewrite host, uri, headers, and more before requests reach upstreams, or customize responses returned to clients.
  • Circuit breaker: Intelligently tracks unhealthy upstream services to prevent fault propagation.
  • Cross-platform deployment: No vendor lock-in. Runs anywhere from bare metal to Kubernetes, with ARM64 support.
  • Kubernetes Ingress: Can be used as a Kubernetes ingress controller to handle cluster ingress traffic.

Releases

Notify me

Initial · Apache APISIX

Q&A

What is the underlying tech stack of Apache APISIX?AI-generated
APISIX is written in Lua and built on top of NGINX and etcd.
How do I quickly install and start APISIX?AI-generated
In a Docker-enabled environment, run curl -sL https://run.api7.ai/apisix/quickstart | sh to start APISIX (listening on port 9080) along with its etcd configuration store in one step.
Can APISIX be used as an AI gateway?AI-generated
Yes. APISIX provides AI proxy capabilities through its flexible plugin system, routing traffic uniformly to different LLM providers with cross-LLM load balancing, retries, fallbacks, and token-based rate limiting.
What is the mcp-bridge plugin?AI-generated
mcp-bridge is an APISIX plugin that seamlessly converts stdio-based MCP servers into scalable HTTP SSE services.
Which network protocols does APISIX support?AI-generated
It supports HTTP/HTTPS, TCP/UDP, Dubbo, MQTT (3.1.x and 5.0), gRPC, gRPC Web, WebSocket, Proxy Protocol, HTTP/3 (QUIC), and more.
Do I need to restart APISIX to update configurations or plugins?AI-generated
No. APISIX supports hot reloading of configurations and plugins, so changes take effect immediately without a restart.
Does APISIX support Kubernetes?AI-generated
Yes. APISIX is cloud-native, runs anywhere from bare metal to Kubernetes, and can serve as a Kubernetes ingress controller.
More
View repo
Bookmark

This product has no owner yet

It was listed by the platform or its author left. If it is your work, prove your relationship to become the owner.

Claim this product

Alternatives

The product content on this page was translated by AI from the author's original.