Docs
API reference
Responses

Responses

POSThttps://api.airoutly.com/v1/responses

OpenAI's newer response API, used by Codex CLI and recent GPT models. Requests and responses follow the OpenAI format exactly.

Authentication

Send your key as Authorization: Bearer $AIROUTLY_API_KEY, or pass api_key to an OpenAI SDK.

Request body

Parameter	Description
`model`stringrequired	A model that lists Resp in the Endpoints column of the models page, such as `gpt-5`.
`input`string \| arrayrequired	A prompt string, or an array of input items (messages, images, tool results).
`instructions`string	System-level instructions for this response.
`stream`boolean	Stream events as they are generated.
`max_output_tokens`integer	Upper limit on generated tokens, including reasoning tokens.
`reasoning`object	For example `{"effort": "low"}` on reasoning models.
`tools`array	Functions the model may call.

Other OpenAI parameters are passed through to the model.

Example request

curl https://api.airoutly.com/v1/responses \
  -H "Authorization: Bearer $AIROUTLY_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"model": "gpt-5", "input": "Write a haiku about latency."}'

import os
from openai import OpenAI

client = OpenAI(base_url="https://api.airoutly.com/v1", api_key=os.environ["AIROUTLY_API_KEY"])
resp = client.responses.create(model="gpt-5", input="Write a haiku about latency.")
print(resp.output_text)

import OpenAI from "openai";

const client = new OpenAI({ baseURL: "https://api.airoutly.com/v1", apiKey: process.env.AIROUTLY_API_KEY });
const resp = await client.responses.create({ model: "gpt-5", input: "Write a haiku about latency." });
console.log(resp.output_text);

Response

200 OK

{
  "id": "resp_123",
  "object": "response",
  "model": "gpt-5",
  "status": "completed",
  "output": [{
    "type": "message",
    "role": "assistant",
    "content": [{ "type": "output_text", "text": "Packets in the dark…" }]
  }],
  "usage": { "input_tokens": 14, "output_tokens": 19, "total_tokens": 33 }
}

You are billed for usage at the model’s price on the models page. Reasoning tokens count as output.

Streaming

With "stream": true the reply arrives as server-sent events (response.output_text.delta, …) ending with response.completed, which carries the final usage.

Stateless usePass the full conversation in input each time rather than relying on stored responses.