Responses

Responses provides a unified input and output structure for Codex, modern agents, tools, and multimodal workflows. Confirm that the model and group support this protocol.

curl

# Responses uses /v1/responses and the same Bearer authentication.
curl https://sprelaytoken.com/v1/responses \
  -H "Authorization: Bearer $SPRELAY_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gpt-5.6-sol",
    "input": "Reply only: Responses connection successful"
  }'

model must support Responses, and input is the request content. Preserve the JSON quotes and commas when editing.

Python SDK

# Import the SDK and read the key from the environment.
import os
from openai import OpenAI

# Keep the base URL at /v1; the SDK appends /responses.
client = OpenAI(
    api_key=os.environ["SPRELAY_API_KEY"],
    base_url="https://sprelaytoken.com/v1",
)

# A plain string is the easiest input for the first test.
response = client.responses.create(
    model="gpt-5.6-sol",
    input="Give three reasons not to put API keys in frontend code.",
)

# output_text is the SDK's convenient aggregated text field.
print(response.output_text)

JavaScript SDK

// Import the SDK; read the key from the Node.js environment.
import OpenAI from 'openai'

const client = new OpenAI({
  apiKey: process.env.SPRELAY_API_KEY,
  baseURL: 'https://sprelaytoken.com/v1',
})

// The SDK sends this object to POST /v1/responses.
const response = await client.responses.create({
  model: 'gpt-5.6-sol',
  input: 'Give three reasons not to put API keys in frontend code.',
})

// Print only the aggregated text.
console.log(response.output_text)

Chat Completions usually returns choices; Responses returns output, while SDKs commonly expose output_text. Do not reuse a Chat Completions response parser unchanged. A 404 usually means /v1 or /responses was appended twice.