OpenAI-compatible API checker
Check whether an endpoint really behaves like OpenAI.
Verify the practical parts of OpenAI compatibility: model listing, chat completions, response shape, streaming chunks, and whether a claimed model family behaves as expected.
Make compatibility visible
Some gateways accept OpenAI-style payloads but fail streaming, return odd chunks, or expose model names that do not behave like the expected model family.
What Compatibility Means
Compatibility means the endpoint accepts OpenAI-style request payloads, returns choices in a familiar shape, supports /v1/models, and can stream server-sent events when stream is set to true.
Who Should Use It
This page is useful for AI relay users, proxy buyers, API gateway builders, and developers who want to switch providers without rewriting their OpenAI client code.
Compatibility vs Speed
A compatible endpoint is not always a good endpoint. It can pass the format test and still have slow TTFT, unstable streaming, or long total response time. That is why the main tool checks latency alongside compatibility.
FAQ
What is an OpenAI-compatible API?
It is an API that follows common OpenAI endpoint paths and payload shapes, so OpenAI client libraries can often connect to it.
Does compatible mean official OpenAI?
No. Many relays, proxies, and gateways are compatible with the OpenAI API shape but are not official OpenAI services.
Why does chat work but streaming fail?
Some providers implement normal chat completions but do not correctly support SSE streaming chunks.
Can compatibility pass while latency is slow?
Yes. Compatibility checks the API shape. Latency checks the user experience.
Check compatibility before switching providers.
Test OpenAI-style paths, chat completions, streaming chunks, and model behavior in one place.