MCP Session Memory Bridge

MCP Session Memory Bridge

Provides cross-session memory for AI agents by maintaining a rolling 24-hour event stream and searchable daily archives to prevent context loss. It enables agents to record, query, and retrieve historical events and decisions through a structured markdown-based workspace.

Category
访问服务器

README

MCP Session Memory Bridge

A Model Context Protocol (MCP) server that provides cross-session memory capabilities for AI agents. This solves the common problem of "session amnesia" where agents lose context between sessions.

Features

  • 24-Hour Event Stream: Maintains a rolling RECENT_EVENTS.md file with events from the last 24 hours
  • Daily Memory Archives: Automatically creates daily memory files for long-term storage
  • Session Context Preservation: Loads relevant context when starting a new session
  • Event Tagging: Tag events with categories like [Decision], [Action], [Context], [User]
  • Searchable History: Full-text search across all memory files

Installation

npm install mcp-session-memory

Usage

Start the MCP Server

npx mcp-session-memory --workspace /path/to/workspace

In Your Agent Code

import { Client } from "@modelcontextprotocol/sdk/client/index.js";
import { StdioClientTransport } from "@modelcontextprotocol/sdk/client/stdio.js";

const transport = new StdioClientTransport({
  command: "npx",
  args: ["mcp-session-memory", "--workspace", "./workspace"]
});

const client = new Client({
  name: "my-agent",
  version: "1.0.0"
}, {
  capabilities: {}
});

await client.connect(transport);

// Record an event
await client.callTool("memory_record", {
  category: "Decision",
  content: "Approved new skill generation workflow",
  tags: ["workflow", "approval"]
});

// Query memory
const results = await client.callTool("memory_query", {
  query: "skill workflow decisions",
  limit: 5
});

// Get session context
const context = await client.callTool("session_getContext", {
  maxAge: "24h"
});

Tools

memory_record

Records a new event to the memory stream.

Input:

  • category: "Decision" | "Action" | "Context" | "User" | "System"
  • content: string - The event content
  • tags: string[] - Optional tags for categorization
  • importance: number (1-10) - Importance level

memory_query

Searches across all memory files.

Input:

  • query: string - Search query
  • limit: number - Maximum results (default: 10)
  • categories: string[] - Filter by categories

session_getContext

Gets relevant context for starting a new session.

Input:

  • maxAge: string - e.g., "24h", "7d" (default: "24h")
  • limit: number - Maximum events to return

memory_getRecent

Gets recent events from the stream.

Input:

  • hours: number - Hours to look back (default: 24)
  • limit: number - Maximum events to return

File Structure

workspace/
├── RECENT_EVENTS.md      # 24-hour rolling event stream
├── memories/
│   ├── 2026-03-15.md    # Today's memory
│   ├── 2026-03-14.md
│   └── ...
└── .memory-config.json   # Configuration

RECENT_EVENTS.md Format

# Recent Events - Last 24 Hours

## 2026-03-15 10:30:00
- [Decision] Approved new skill generation workflow
- [Tags] workflow, approval

## 2026-03-15 09:15:00
- [Action] Completed morning operation for EvoMap
- [Context] Active project: Community Platform

Configuration

{
  "workspace": "./workspace",
  "retentionDays": 30,
  "maxEventsPerFile": 100,
  "categories": ["Decision", "Action", "Context", "User", "System"]
}

License

MIT

推荐服务器

Baidu Map

Baidu Map

百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。

官方
精选
JavaScript
Playwright MCP Server

Playwright MCP Server

一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。

官方
精选
TypeScript
Audiense Insights MCP Server

Audiense Insights MCP Server

通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。

官方
精选
本地
TypeScript
Magic Component Platform (MCP)

Magic Component Platform (MCP)

一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。

官方
精选
本地
TypeScript
VeyraX

VeyraX

一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。

官方
精选
本地
Kagi MCP Server

Kagi MCP Server

一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。

官方
精选
Python
graphlit-mcp-server

graphlit-mcp-server

模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。

官方
精选
TypeScript
Exa MCP Server

Exa MCP Server

模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。

官方
精选
mcp-server-qdrant

mcp-server-qdrant

这个仓库展示了如何为向量搜索引擎 Qdrant 创建一个 MCP (Managed Control Plane) 服务器的示例。

官方
精选
e2b-mcp-server

e2b-mcp-server

使用 MCP 通过 e2b 运行代码。

官方
精选