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

# Responses API

> POST /v1/responses, OpenAI Responses API-compatible.

`POST https://asklethe.ai/v1/responses` accepts the same request shape as OpenAI's Responses API (`input` instead of `messages`, structured output items, etc.). Use this if your client already targets the Responses API rather than Chat Completions.

<CodeGroup>
  ```bash cURL theme={null}
  curl https://asklethe.ai/v1/responses \
    -H "Authorization: Bearer $LETHE_API_KEY" \
    -H "Content-Type: application/json" \
    -d '{
      "model": "lethe",
      "input": "Say hi in five words."
    }'
  ```
</CodeGroup>

## When to use this vs. Chat Completions

* Use [Chat Completions](/api-reference/chat-completions) if your tooling (LangChain, CrewAI, most SDKs) expects the `messages` array shape.
* Use Responses if you're already on OpenAI's newer Responses API and want to keep that request/response shape.

Both endpoints share the same authentication, quota, model resolution (`lethe` / `lethe-vision`), and error format. See [Authentication](/api-reference/authentication), [Rate Limits](/api-reference/rate-limits), and [Errors](/api-reference/errors).

Vision input (`input_image` / `input_video` items) requires `model: "lethe-vision"`, same rule as Chat Completions.