Responses

Use it when

• Your app already uses the OpenAI Responses API shape.
• You send input and optional instructions instead of chat-only messages.
• You want one route for text output and tool-capable response flows.

Code examples

curl https://api.prismaticapi.com/v1/responses \
  -H "Authorization: Bearer $PRISMATIC_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "claude-sonnet-4-6",
    "input": "Summarize the changelog in four bullets.",
    "max_output_tokens": 400
  }'

Parameters

• model (string, required) — Public model ID returned by GET /v1/models.
• input (string or array, required) — User input in Responses format.
• instructions (string, optional) — Developer or system instructions.
• max_output_tokens (integer, optional) — Maximum generated output tokens.
• stream (boolean, optional) — Enables streaming when supported.
• tools and tool_choice are accepted when supported by the selected model.

Notes

• The endpoint uses the same billing decision layer as Chat Completions.
• Unsupported shapes return invalid_request with validation details.