OpenAIResponsesRequest
A request to the OpenAI Responses API.
System-managed fields (streaming, conversation chaining on the server side) are intentionally not exposed for client-side processing.
Constructors
Properties
Whether to run the response as a background job. Defaults to false.
The conversation input: plain text or a list of input items.
The system/developer prompt for the model. null uses provider default.
An upper bound for the number of tokens generated. null uses provider default.
An upper bound for the number of tool calls generated. null uses provider default.
Up to 16 key-value pairs attached to the request. null when not set.
Whether the model can run tool calls in parallel. null uses provider default.
The ID of the previous response to chain from (multi-turn). null starts a new conversation.
A cache key used to route the request for prompt caching. null when not set.
The reasoning configuration for reasoning models. null uses provider default.
A stable identifier used for safety/abuse tracking. null when not set.
The processing tier used to serve the request. null uses provider default.
Sampling temperature (0.0–2.0). null uses provider default.
Text output configuration (format, verbosity). null uses provider default.
Which (if any) tool the model is forced to call. null uses provider default.
How registered tools are executed by the client. Defaults to auto-execute.
The tools available to the model. Defaults to an empty list.
The truncation strategy for the context window. null uses provider default.