记忆系统
How Suzent remembers things across conversations, where it keeps them, and how to change what it knows.
How memory is written, consolidated, indexed, and read. Companions: Consolidation for how the dream behaves from the outside, Internals for the class-by-class reference. Three tier
Memory ConfigurationMost people never open this page. Memory is on out of the box, and the two settings you are most likely to want — whether it runs at all, and which models it uses — are in Settings
How Consolidation WorksSuzent captures fast and tidies up later. Every conversation appends raw facts to a daily log; a background pass called the dream turns those logs into organised notebook pages. Th
Memory InternalsThis page describes the implementation-level structure of Suzent memory as it exists today. Layers Layer Responsibility --- --- Agent prompt and tools memory_search , block_update
NotebookYour agent keeps a second kind of memory: a notebook , an Obsidian-compatible folder of Markdown pages it researches, writes, and tidies itself. Conversation memory records facts a
记忆系统统一的记忆-会话-智能体架构,具备人类可读的 Markdown 持久化、JSONL 会话记录和可检查的智能体状态。 概述 Markdown 记忆 ( /shared/memory/ ):人类可读的真实来源——每日日志和精选的 MEMORY.md,智能体可通过文件工具直接访问。 LanceDB 搜索索引 ( .suzent/memory/ ):对提取的事实进行
Upgrade NotesThe upgrade guide has moved to Upgrading Suzent. It covers migration steps and changes affecting existing installations.