Create response

Creates a model response for the given input. This is the OpenAI Responses API. Supports instructions (system message equivalent), streaming, tool calling, JSON mode, and multi-turn conversations. When present, reasoning is returned as a structured output item with type: "reasoning", separate from the message output.

POST /responses

cURL

Basic

curl --request POST \
  --url https://ai.eu.corti.app/v1/responses \
  --header 'Authorization: Bearer <token>' \
  --header 'Content-Type: application/json' \
  --data '
{
  "model": "corti-s1",
  "input": [
    {
      "role": "user",
      "content": [
        {
          "type": "input_text",
          "text": "Count from 1 to 5"
        }
      ]
    }
  ]
}
'

200

{
  "id": "<string>",
  "object": "response",
  "created_at": 123,
  "model": "<string>",
  "incomplete_details": {
    "reason": "max_output_tokens"
  },
  "instructions": "<string>",
  "output": [
    {
      "id": "<string>",
      "type": "message",
      "role": "assistant",
      "status": "completed",
      "phase": "<string>",
      "content": [
        {
          "text": "<string>",
          "type": "output_text",
          "annotations": [
            "<unknown>"
          ],
          "logprobs": {}
        }
      ]
    }
  ],
  "temperature": 123,
  "top_p": 123,
  "tool_choice": "<string>",
  "tools": "<array>",
  "max_output_tokens": 123,
  "previous_response_id": "<string>",
  "usage": {
    "input_tokens": 123,
    "output_tokens": 123,
    "total_tokens": 123,
    "input_tokens_details": {
      "cached_tokens": 123,
      "input_tokens_per_turn": [
        "<unknown>"
      ],
      "cached_tokens_per_turn": [
        "<unknown>"
      ]
    },
    "output_tokens_details": {
      "reasoning_tokens": 123,
      "tool_output_tokens": 123,
      "output_tokens_per_turn": [
        "<unknown>"
      ],
      "tool_output_tokens_per_turn": [
        "<unknown>"
      ]
    }
  },
  "service_tier": "<string>",
  "truncation": "<string>",
  "background": true,
  "parallel_tool_calls": true,
  "max_tool_calls": 123,
  "text": "<string>",
  "top_logprobs": 123,
  "reasoning": "<string>",
  "prompt": "<string>",
  "metadata": {},
  "user": "<string>",
  "presence_penalty": 123,
  "frequency_penalty": 123,
  "kv_transfer_params": {},
  "input_messages": "<array>",
  "output_messages": "<array>"
}

{
  "error": "<string>",
  "detail": "<string>"
}

{
  "error": "<string>",
  "detail": "<string>"
}

{
  "error": "<string>",
  "detail": "<string>"
}

Authorizations

Authorization · string · header · required

Use a Corti-issued bearer token.

Body

application/json

model · enum<string> · default: corti-s1 · required

Available options: corti-s1, corti-s1-instant, corti-s1-mini, corti-s1-mini-instant

input · object[] · required

Input messages as an array of message objects. Each message has a role and a content array of typed parts.

Show child attributes

instructions · string | null

System-level instructions. Equivalent to a system message in chat completions.

stream · boolean · default: false

If true, returns Server-Sent Events (SSE) with typed events (response.created, response.in_progress, response.output_item.added, response.output_text.delta, response.completed).

temperature · number<float> · default: 1

Required range: 0 <= x <= 1

top_p · number<float> · default: 0.95

max_output_tokens · integer

Maximum output tokens. If exceeded, the response status is incomplete with incomplete_details.reason: "max_output_tokens".

Required range: x >= 1

response_format · object

Show child attributes

tools · object[]

Tools the model may call.

Show child attributes

tool_choice · enum<string> · default: auto

Available options: auto, none

presence_penalty · number<float> · default: 0

frequency_penalty · number<float> · default: 0

Response

200

application/json

Successful response. For non-streaming requests, returns a complete response object. For streaming requests, returns Server-Sent Events (SSE) with typed events.

id · string

Unique response ID (e.g. resp_88d5b93fd2e7b7b1).

object · enum<string>

Available options: response

created_at · integer

Unix timestamp (seconds).

model · string

The model name used for this response.

status · enum<string>

incomplete when max_output_tokens is hit before generation finishes.

Available options: completed, incomplete

incomplete_details · object | null

Show child attributes

instructions · string | null

output · (Message · object | Function Call · object)[]

Array of output items. May include reasoning items with type: "reasoning", message items with type: "message", and function call items with type: "function_call".

Message

Function Call

Show child attributes

temperature · number

top_p · number

tool_choice · string

tools · array

max_output_tokens · integer

previous_response_id · string | null

usage · object

Show child attributes

service_tier · string | null

truncation · string

background · boolean

parallel_tool_calls · boolean

max_tool_calls · integer | null

text · string | null

top_logprobs · integer | null

reasoning · string | null

Reasoning is returned in output[] items.

prompt · string | null

metadata · object | null

user · string | null

presence_penalty · number

frequency_penalty · number

kv_transfer_params · object | null

input_messages · array | null

output_messages · array | null