OpenAI-compatible API

OpenAI's API format has become the de facto standard for how AI applications talk to language models. The format defines how you send messages, what parameters you include (like temperature and max tokens), and how streamed or complete responses come back. Because so many tools, SDKs, and frameworks were built to work with this format, a provider that matches it gains immediate compatibility with the entire OpenAI ecosystem — no custom integration needed.

In practice, switching from OpenAI to an OpenAI-compatible provider often means changing just one line: the base URL the client points to. Providers like Groq, Together AI, Ollama, and vLLM all offer OpenAI-compatible endpoints. This portability is one reason the ecosystem has coalesced around this standard rather than fragmenting into incompatible interfaces.

For builders, the practical benefit is flexibility without lock-in. You can develop against one provider, then swap to another for cost, speed, or model choice reasons, and the rest of your code stays the same. Tools like model routers and AI gateways use this standard to route traffic across multiple providers behind a single unified endpoint.