POST/v1/responses
Responses
POST /v1/responses: create a model response in the OpenAI Responses format, for the models that support it.
La documentation est en anglais.
Creates a model response in the OpenAI Responses format. Works with the models that list /v1/responses in their endpoints. Some of them are only available here.
Base URL: https://api.nexrouting.com/v1 (also https://nexrouting.com/api/v1).
Headers
AuthorizationstringrequiredAuthorization: Bearer <key>(orx-api-key: <key>).Content-Typestringapplication/json. The body is read as JSON either way.
Body
modelstringrequired- Model ID, such as
openai/gpt-6-astra. inputstring | arrayrequired- The prompt: a string, or a list of input items.
instructionsstring- System instructions.
streambooleantrueto receive server-sent events. See Streaming.max_output_tokens, reasoning, tools, …- Passed on as they are, as in the OpenAI API.
The gateway checks model and stream. Everything else in the body goes to the model provider unchanged, in the OpenAI Responses format, so which other parameters work depends on the model. When the provider rejects a parameter, you get 400 invalid_request with its reason.
Example request
curl https://api.nexrouting.com/v1/responses \ -H "Authorization: Bearer $NEXROUTING_API_KEY" \ -H "Content-Type: application/json" \ -d '{ "model": "openai/gpt-6-astra", "input": "Explain routing in one sentence." }'Example response
{ "id": "resp_68a1c0", "object": "response", "created_at": 1795075200, "status": "completed", "model": "openai/gpt-6-astra", "output": [ { "type": "message", "id": "msg_68a1c1", "status": "completed", "role": "assistant", "content": [ { "type": "output_text", "text": "Routing sends each request to the model that should answer it.", "annotations": [] } ] } ], "usage": { "input_tokens": 12, "input_tokens_details": { "cached_tokens": 0 }, "output_tokens": 14, "output_tokens_details": { "reasoning_tokens": 0 }, "total_tokens": 26 }}Errors
OpenAI error shape. All codes are in Errors.
Not available
Retrieving, cancelling or deleting a stored response (/v1/responses/{id} and its sub-paths) returns 404 not_found. A follow-up request with previous_response_id can be served by another route that doesn't know the earlier response, so send the conversation in input instead.