> ## Documentation Index
> Fetch the complete documentation index at: https://docs.openserv.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Choose an API endpoint

> Select the endpoint that matches your client and model.

<Frame>
  <img src="https://mintcdn.com/openserv/wCXyJJTiOX6drWSJ/images/tutorials/choose-an-endpoint.webp?fit=max&auto=format&n=wCXyJJTiOX6drWSJ&q=85&s=9ff6a09db95590048c375c39825850bf" alt="One application request branching into three API endpoint lanes and converging on a model catalog" width="1536" height="1024" data-path="images/tutorials/choose-an-endpoint.webp" />
</Frame>

| If you need                                     | Use                                               |
| ----------------------------------------------- | ------------------------------------------------- |
| One endpoint for any catalog model              | [`/v1/chat/completions`](../api/chat-completions) |
| OpenAI Responses format or reasoning summaries  | [`/v1/responses`](../api/responses)               |
| An Anthropic SDK or Messages-shaped integration | [`/v1/messages`](../api/messages)                 |

`/v1/responses` supports OpenAI-routed models only. For Gemini, Claude, Gemma, Grok, Qwen, or DeepSeek, use Chat Completions unless the [compatibility matrix](../api/compatibility) says otherwise.

The endpoint changes the request and response shape, not the SERV reasoning concept. Use the [parameter map](../sdk-integration#parameter-map) when translating system prompts, token limits, reasoning controls, tools, and usage fields between formats.