Documentation
Documentation menu

Responses

A minimal text request and output handling without mixing protocols.

On this page

Send input

POST https://api.etage.one/v1/responses uses model and input. The example selects gpt-6.1-sol. Set ETAGE_API_KEY securely in your environment beforehand; running it uses credits. Python and TypeScript examples need no separate client library. Do not copy a Chat Completions body into this format without adapting it.

Bash
curl --fail-with-body -sS https://api.etage.one/v1/responses \
  -H "Authorization: Bearer $ETAGE_API_KEY" \
  -H "Content-Type: application/json" \
  --data '{"model":"gpt-6.1-sol","input":"Hello"}'

Read text output items

In a regular JSON response, iterate through output, then content, and select items with type: output_text. Not every output item is a text message. Check the HTTP status and response status: an incomplete or failed result must not be treated as a completed answer. This minimal example does not establish support for tools, continuation by identifier or other optional fields.