gateway

Why we included this project

Teams building on large language models soon find that hooking each provider up directly means juggling different SDKs, rate limits, and response formats. This gateway sits in front of all of that and exposes a single API that reaches over 1,600 models across many providers, so an application can switch vendors or route specific requests without touching its own code. It also bundles guardrails plus supporting infrastructure like caching and observability that most teams would otherwise have to assemble themselves. For developers and platform engineers running production traffic against several LLM providers, it is a reasonable single place to handle routing, cost, and safety policies before you build that plumbing by hand.