"Conversations" issue with "store": false - an issue to be rectified

API
1

I hope this message finds you well, but probably won’t find you at all…

There are many forum posts illustrating this pain point with current behavior:

  • Use "store": false with “conversations” as input, not wishing to persist a response ID
  • and you get new conversation turns silently never added to the conversation object.

This simply makes no sense.

Use of conversations already implies consent to persistent “forever” data retention of those objects and user data. The implied contract there: you can add messages directly to a conversation, no storage parameter needed to block that, and you have implicitly acknowledged that zero-data retention is not a goal when employing this “conversations” endpoint on the Responses API.

Sending the signal that you do not want response IDs logged and exposed in parallel (also with no mechanism to list them by API endpoint and clean up the mess) should work and function to stop logging. It should mean NO RESPONSE ID. It should NOT mean not NO WORKING CONVERSATION to fail silently against expected use.

Action needed: store the newest input and AI model output as normal and as would be expected by a reasonable person when store:false is expressed, simply to turn off the response ID storage.

@wilkes likes fixing the conversations API.

Separate action: make the “previous_response_id” field nullable on the responses endpoint. Issue: The API endpoint is intercepting any use of “previous_response_id” parameter along with “conversations” in the same call, including attempts at None/null formed by a default shape, and returning errors.

OpenAI is even returning previous_response_id in the response.completed event object, breaking such an input validation “schema” by example. See “conversation” and “previous_response_id” right now:

    "object": "response",
    "created_at": 1758061543,
    "status": "completed",
    "background": false,
    "conversation": {
      "id": "conv_68xxx"
    },
    "error": null,
    "incomplete_details": null,
    "instructions": "xxxt.",
    "max_output_tokens": null,
    "max_tool_calls": null,
    "model": "gpt-5-mini-2025-08-07",
    "output": [
      {
        "id": "rs_052xxx",
        "type": "reasoning",
        "summary": []
      },
      {
        "id": "msg_0520xxx",
        "type": "message",
        "status": "completed",
        "content": [
          {
            "type": "output_text",
            "annotations": [],
            "logprobs": [],
            "text": "xxx"
          }
        ],
        "role": "assistant"
      }
    ],
    "parallel_tool_calls": true,
    "previous_response_id": null,
    "prompt_cache_key": null,
    "reasoning": {
      "effort": "medium",
      "summary": null
    },
Conversation List Contain Duplicate User Message