LycheeMem

概述

LycheeMemory 是一个专为 LLM 智能体设计的紧凑记忆框架，通过结构化组织、轻量级整合和自适应检索提供高效的对话记忆功能。它作为 HTTP 服务器实现 MCP 协议支持，兼容任何 MCP 客户端。该框架使用 LanceDB 进行向量存储，支持多种重排模型以提高记忆检索准确性。项目维护活跃，最近的更新包括 Transformer 重排器 v0 和视觉记忆模块。

试试问 AI

装完之后，这里有 3 个你可以让 AI 做的事：

• lychee_memory_smart_search

  Default long-term memory retrieval entry point

  试试:Search for information about previous conversations试试:Find relevant memories about topic X
• lychee_memory_search

  General memory search function

  试试:Search for specific memories试试:Find all memories related to Y
• lychee_memory_append_turn

  Add a new turn to the memory

  试试:Append this user message to memory试试:Save this assistant response to the conversation
• lychee_memory_synthesize

  Synthesize information from memory

  试试:Create a summary of recent conversations试试:Generate insights from my memory
• lychee_memory_consolidate

  Consolidate memory turns

  试试:Consolidate current session memories试试:Merge related memory entries

可对比工具

安装

pip install lycheemem

pip install "lycheemem[rerank]"

lycheemem-cli

{
  "mcpServers": {
    "lycheememory": {
      "command": "python",
      "args": ["-m", "lycheemem", "server"],
      "env": {}
    }
  }
}

LycheeMem 对比