Responses

Use the Responses route for response-style input and tool-capable model calls.

POST /v1/responses

Request

curl https://api.voxvey.com/v1/responses \
  -H "Authorization: Bearer $VOXVEY_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "openai/gpt-5.5",
    "instructions": "You are a concise API documentation assistant. Return only the requested answer.",
    "input": "Turn this into three short action items: get an access token, call the models endpoint, send a chat completion request.",
    "stream": true
  }'

Example streamed response

Streaming responses are sent as server-sent events. Concatenate response.output_text.delta values to build the final text.

data: {"type":"response.created","response":{"id":"resp_...","object":"response","status":"in_progress","model":"gpt-5.5"}}

data: {"type":"response.output_text.delta","delta":"1. Get an access token.  \n"}

data: {"type":"response.output_text.delta","delta":"2. Call the models endpoint to verify API access.  \n"}

data: {"type":"response.output_text.delta","delta":"3. Send a chat completion request."}

data: {"type":"response.completed","response":{"id":"resp_...","object":"response","status":"completed","model":"gpt-5.5"}}

data: [DONE]

Final text:

1. Get an access token.
2. Call the models endpoint to verify API access.
3. Send a chat completion request.

Required fields

Field	Type	Notes
model	string	Provider-prefixed model ID
input	string or array	Response input

Notes

• BytePlus response calls may use ModelArk response fields such as input, instructions, tools, thinking, and include.
• The gateway settles successful calls from usage metadata when available.