{
  "id": 7222769,
  "title": "Ollama's Responses API accepts previous_response_id, returns 200, and forgets the whole conversation",
  "url": "https://urgent.news/2026/09/14/ollamas-responses-api-accepts-previous-response-id-returns-200-and",
  "topic": "tech",
  "section": "Tech",
  "published": "2026-09-14T02:10:04.000Z",
  "source": {
    "name": "Dev.to",
    "slug": "dev-to",
    "url": "https://dev.to/homelabpm/ollamas-responses-api-accepts-previousresponseid-returns-200-and-forgets-the-whole-conversation-1kp8"
  },
  "original_language": "en",
  "account": "Ollama's Responses API accepts a previous_response_id in requests, returns a 200 status, but forgets the entire conversation context. The model treats each request as a new conversation, regardless of the previous_response_id sent. This is due to the request struct lacking a previous_response_id field, which gets discarded during JSON decoding. The server's response includes a null previous_response_id field, indicating the feature is not supported. To resolve this, clients should send the full conversation history with each request to ensure the model has the necessary context.",
  "summary": "TL;DR : send previous_response_id to Ollama's /v1/responses and it is accepted, the reply is HTTP 200 with \"status\": \"completed\" and \"error\": null , and the model has never seen the previous turn. A request that carries previous_response_id costs the same number of input tokens as a request that carries no history — 41 and 41 on this box — and produces the same kind of answer. Sending the full…",
  "key_points": [
    "Ollama's Responses API accepts previousresponseid in requests.",
    "API returns 200 status but forgets entire conversation context.",
    "Previousresponseid field discarded during JSON decoding."
  ],
  "editors_take": "This oversight means clients must adapt by sending full conversation histories with each request, rather than relying on the API to maintain context.",
  "illustration": null,
  "coverage": {
    "outlets": 1,
    "also_reported_by": []
  },
  "ai_generated": true,
  "disclaimer": "Summaries, key points and the editor’s take are written by software from other outlets’ reporting and may contain errors — always check the linked original."
}