<?xml version="1.0" encoding="utf-8" standalone="yes"?><rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>Persistence on PlumePHP</title><link>https://plumephp.com/tags/persistence/</link><description>Recent content in Persistence on PlumePHP</description><generator>Hugo</generator><language>zh-CN</language><lastBuildDate>Sat, 26 Sep 2026 00:00:00 +0800</lastBuildDate><atom:link href="https://plumephp.com/tags/persistence/index.xml" rel="self" type="application/rss+xml"/><item><title>Agent 记忆系统实战：从上下文窗口到长期记忆的持久化架构</title><link>https://plumephp.com/llm-agent-memory-persistence/</link><pubDate>Sat, 26 Sep 2026 00:00:00 +0800</pubDate><guid>https://plumephp.com/llm-agent-memory-persistence/</guid><description>&lt;p&gt;LLM 本身没有记忆——每次调用都从空白上下文开始。但真实的智能应用需要跨会话记住用户偏好、任务进度与历史事实。Agent 记忆系统正是为了解决&amp;quot;上下文是易失的、知识是永恒的&amp;quot;这一矛盾而生的工程领域。本指南从记忆的类型学出发，系统覆盖短期工作记忆、长期语义记忆、情景记忆与程序性记忆的落地实现，并给出向量库、SQLite、文件系统三种存储方案的完整代码。&lt;/p&gt;</description></item></channel></rss>