{
  "id": 9039956,
  "title": "06 — Chat Works. Does the Agent Actually Retrieve Memory?",
  "url": "https://urgent.news/2026/09/22/06-chat-works-does-the-agent-actually-retrieve-memory",
  "topic": "ai",
  "section": "AI",
  "published": "2026-09-22T00:03:16.000Z",
  "source": {
    "name": "Dev.to",
    "slug": "dev-to",
    "url": "https://dev.to/liuyuyan6100/06-chat-works-does-the-agent-actually-retrieve-memory-21hf"
  },
  "original_language": "en",
  "account": "The article explores whether an agent can actually retrieve memory, rather than just present information in context during a conversation. Initial tests showed that despite memory content appearing in the conversation context, the agent's tool execution environment might not be able to reach the memory addresses provided.\n\nTwo types of pathways exist for memory retrieval - passive injection, where the server side reads memory and includes it in context, and active retrieval using endpoints supplied in tool prompts. The article found that while passive injection continued working, active retrieval failed, even though the server's read path was functional.\n\nKey factors were a missing configuration field called injection.externalGatewayUrl, which defaults to selecting an internal IP address when none is provided. This resulted in retrieval requests timing out or receiving no response, as the client tool execution environment was on an external network separate from the internal one where memory was stored.\n\nThe article stresses that merely seeing memory content in context does not prove retrieval functionality. Rather, it needs testing if the client can actually reach the memory address through proper configuration. Other probing results like HTTP authentication errors (401) or no response (000) also needed separate investigation, as these indicated issues in network connectivity and routing rather than memory storage itself.\n\nIn summary, while passive injection provided the conversation context with memory content, active retrieval proved problematic due to network configuration issues between the model's tool executor and the actual memory storage location. Fixing the missing configuration parameter resolved the active retrieval failure, but other independent issues related to knowledge resource URLs still needed separate verification.",
  "summary": "The images in this article are explanatory diagrams with Chinese labels, not screenshots of test results. Each caption summarizes the diagram in English. I wanted to make projects easier to resume across conversations and agents. Before testing whether the next agent could pick up where the previous one stopped, I checked something more basic: could it actively retrieve memory? In the Docker…",
  "key_points": [
    "Initial tests showed memory content in conversation context",
    "Active retrieval failed despite functional server read path",
    "Missing configuration field caused active retrieval issues"
  ],
  "editors_take": "Fixing a single missing configuration field resolved active memory retrieval issues, highlighting that apparent memory access can mask underlying network configuration problems that must be addressed separately.",
  "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."
}