Create model response

Create a model response using the same interface as the OpenAI Responses API. Supports both SSE (Server-Sent Events) streaming (when stream: true) and non-streaming modes. Request and response formats follow the OpenAI Create a model response API.

Tools: Server-side tools registered in the gateway are merged with client tools. In non-streaming mode they are auto-executed in a multi-step loop; in streaming mode the model's tool calls are streamed to the client (no auto-execution).

Recent Requests
Log in to see full request history
TimeStatusUser Agent
Retrieving recent requests…
LoadingLoading…
Body Params

Request body follows the OpenAI Responses API format.

Request format follows OpenAI Responses API (create response). Required: model, input. All other OpenAI fields pass through.

string
required

Model ID in <provider>/<model> format. Use /api/v1/model/info to discover available models.

required

Input to the model: a plain text string or an array of structured input items. See OpenAI input items reference.

string

System instructions for the model.

boolean
Defaults to false

If true, response is sent as Server-Sent Events (text/event-stream).

number
0 to 2

Sampling temperature.

number
0 to 1

Nucleus sampling parameter.

integer
≥ 1

Maximum number of output tokens to generate.

tools
array of objects

List of tool definitions the model may call. See OpenAI Responses API for the supported tool schema.

tools
string

Controls tool selection (auto, none, required, or a specific tool object).

string

ID of a previous response to continue a multi-turn conversation.

Headers
string
enum
Defaults to application/json

Generated from available response content types

Allowed:
Responses

Language
Credentials
Bearer
LoadingLoading…
Response
Click Try It! to start a request and see the response here! Or choose an example:
application/json
text/event-stream