Endpoint Details
Review how to make this request and handle its possible responses.
https://integrate.api.nvidia.com/v1/responsesCreates a response using OpenAI's Responses API format. Supports reasoning models with stateful conversation management, tool calls, and advanced reasoning capabilities.
Request Body
Edit the request fields below.
Request details
Parameters10 Parameters
modelrequiredThe model to use for generating the response.
inputrequiredThe input messages for the conversation.
input[].typerequiredThe type of input item.
messageinput[].rolerequiredThe role of the message author.
user · assistant · systeminput[].contentrequiredThe content of the message.
reasoning_effortoptionalControls the effort level for reasoning in reasoning-capable models.
mediumValid values: low · medium · highmax_output_tokensoptionalThe maximum number of output tokens to generate. Reasoning tokens count toward this budget.
16384Minimum: 1Maximum: 32768temperatureoptionalControls randomness in the response generation.
1Minimum: 0Maximum: 2top_poptionalControls nucleus sampling probability mass.
1Minimum: 0Maximum: 1streamoptionalWhether to stream the response as server-sent events.
falsetoolsoptionalA list of tools the model may call.
instructionsoptionalSystem instructions for the conversation.
metadataoptionalAdditional metadata for the request.
Headers3 Headers
AuthorizationrequiredFormat: Bearer $NVIDIA_API_KEY
Content-TyperequiredAcceptoptional