layer-memory
A local-first, Markdown-native AI agent layered memory system that provides MCP tools for storing, recalling, exporting, and importing memories with types like working, persona, and fact.
README
Layer Memory
本地优先、Markdown 原生的 AI 代理分层记忆系统。基于 Midway.js 的 MCP 服务,SQLite 存储,pnpm monorepo。
项目架构
layer-memory/
├── packages/
│ └── core/ # 记忆核心库(框架无关)
│ └── src/store/ # 存储层抽象、SQLite 实现、类型定义
├── apps/
│ ├── mcp-server/ # MCP 服务端(Midway.js 4.0)
│ │ └── src/tools/ # 4 个 MCP Tools
├── docs/ # 文档站(dumi 2.x)
├── skill/ # AI Agent skill 定义
├── .husky/ # Git hooks
├── .changeset/ # 版本管理
├── docker-compose.yml
├── Dockerfile
└── pnpm-workspace.yaml
MCP Tools
| 工具 | 说明 | 参数 |
|---|---|---|
memory_store |
存储一条记忆 | key, value, type(可选), tags(可选) |
memory_recall |
搜索已存储的记忆 | query, type(可选), tags(可选), limit(可选) |
memory_export |
导出所有记忆为 JSON | outputPath |
memory_import |
从 JSON 文件导入记忆 | inputPath |
记忆类型: working (工作记忆) / persona (人格画像) / fact (事实知识)
快速开始
前置要求
- Node.js >= 20
- pnpm 10.x(安装:
corepack enable && corepack prepare pnpm@10.15.0 --activate)
安装
git clone https://github.com/Dante926/layer-memory.git
cd layer-memory
pnpm install
本地开发
pnpm dev
服务启动在 http://localhost:3000,MCP endpoint 在 /mcp。
构建
pnpm build
所有包编译到各自 dist/ 目录。
文档站
pnpm docs:dev
Docker 部署
启动服务
docker compose up -d
服务运行在 http://localhost:3000。
数据持久化
记忆数据存储在 Docker volume layer-memory-data 中,容器重启/更新不会丢失。
重新构建
docker compose build
docker compose up -d
连接到 AI Agent
Claude Code
在 ~/.claude/settings.json 中配置 MCP:
{
"mcpServers": {
"layer-memory": {
"type": "url",
"url": "http://localhost:3000/mcp"
}
}
}
Codex / Cursor
在项目的 .cursor/mcp.json 中配置:
{
"mcpServers": {
"layer-memory": {
"type": "url",
"url": "http://localhost:3000/mcp"
}
}
}
使用示例
存储记忆:
调用 memory_store(
key="架构决策",
value="选择 Midway.js 4.0 MCP + SQLite",
type="fact",
tags="项目:layer-memory"
)
搜索记忆:
调用 memory_recall(query="架构决策", limit=5)
导出备份:
调用 memory_export(outputPath="backup.json")
版本管理
本项目使用 Changeset 管理版本:
pnpm changeset # 创建变更记录
pnpm version # 更新版本号
pnpm publish # 发布
开发
提交规范
使用 Conventional Commits:
feat:新功能fix:修复chore:工程化docs:文档
代码格式化
pnpm format
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.
E2B
Using MCP to run code via e2b.
Neon Database
MCP server for interacting with Neon Management API and databases
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.