Sign in
Model response
POST
/responses
Native OpenAI Responses API endpoint, compatible with the OpenAI SDK's client.responses.create. Just point the SDK's baseURL at SeaWhale AI.
Only OpenAI models are supported for now (e.g. gpt-6-luna, gpt-5). For models from Claude, Gemini, DeepSeek and other vendors, use the "Chat completion" endpoint instead.
Key differences from "Chat completion":
- Uses
inputinstead ofmessages - Uses
instructionsfor the system prompt - Uses
max_output_tokensinstead ofmax_tokens - Returns an
outputarray instead ofchoices; usage fields areinput_tokens/output_tokens
Other official OpenAI Responses parameters not listed below are passed through to the model as-is. Only the create endpoint is available; retrieving or deleting stored responses is not supported.
Request Body
application/json
JSON "model": "gpt-6-luna", "input": "Hello, please introduce yourself"
{
}
Responses
The model response. When stream is true, an SSE (text/event-stream) event stream is returned, each line formatted as data: {event JSON}. Event types match the official OpenAI ones (e.g. response.output_text.delta, response.completed), and the stream ends with data: [DONE].
application/json
JSON "id": "resp_67ccd2bed1ec8190b14f964abc0542670bb6a6b452d3795b", "object": "response", "created_at": 1741476542, "status": "completed", "model": "gpt-6-luna", "output": [ { "type": "message", "id": "msg_67ccd2bf17f0819081ff3bb2cf6508e60bb6a6b452d3795b", "status": "completed", "role": "assistant", "content": [ { "type": "output_text", "text": "Hello! I am an AI assistant that can help you answer questions, write and code.", "annotations": [ ] } ] } ], "usage": { "input_tokens": 14, "input_tokens_details": { "cached_tokens": 0 }, "output_tokens": 22, "output_tokens_details": { "reasoning_tokens": 0 }, "total_tokens": 36 }
{
}