personal-mcp-context

personal-mcp-context

A Model Context Protocol server that stores personal context as structured Markdown files with a knowledge graph, enabling AI agents to query identity, devices, servers, and projects on-demand.

Category
访问服务器

README

Personal MCP Context Registry

A Model Context Protocol (MCP) Server to store and serve your personal context — identity, devices, servers, and projects — as structured Markdown files with a knowledge graph, accessible on-demand by AI Agents (Claude Desktop, Cursor, OpenCode, Antigravity AGY).


Why?

Instead of pasting all your personal info into every AI chat session (wasting tokens and context window), this server lets any MCP-compatible AI agent query only what it needs, when it needs it. Think of it as a personal knowledge base with graph traversal, fuzzy search, and an audit log.


Features

  • 🗂 Graph-backed Storage — Markdown files (personal.md, devices.md, servers.md, etc.) linked via semantic relations in index.json
  • 🔍 Fuzzy Search — Fast in-memory search across all context files via Fuse.js
  • 🔗 Graph Traversal — Query related nodes at configurable depth
  • 🛡 Safe Writes — Every write auto-backs up the existing file before overwriting; rollback on validation failure
  • 📋 Audit Log — JSON-lines access log with auto-purge of entries older than 30 days
  • Integrity Validator — Detects broken links, orphan nodes, and missing files
  • 🐳 Docker-ready — Bind-mount your data from ~/context-data on host, bound exclusively to 127.0.0.1:3031
  • 🗂 Multi-namespace — Separate namespaces for personal, work, project-x contexts

Available MCP Tools

Tool Description
get_context_nodes List all nodes and links in the graph
read_context_file Read full markdown content of a node
search_context Fuzzy search across all context files
get_related_nodes Traverse graph relations from a node
create_context_node Create a new node with markdown content
update_context_file Update existing node content (overwrite / append)
delete_context_node Soft-delete node (moves file to .backup/)
add_context_relation Link two nodes with a named relation
remove_context_relation Remove a link between two nodes
validate_context Run integrity check on graph & files
backup_context Full backup of namespace to .backup/
get_access_log Read the audit access log

Tech Stack

Layer Technology
Runtime Node.js 20 + TypeScript
Transport Streamable HTTP (/mcp endpoint)
Storage Markdown files + JSON graph
Search Fuse.js (in-memory fuzzy)
Auth None (localhost only for MVP)
Deployment Docker Compose

Quick Start

Prerequisites

  • Docker & Docker Compose installed

1. Clone & Create context data directory

git clone https://github.com/your-username/personal-mcp-context.git
cd personal-mcp-context
mkdir -p ~/context-data/personal

2. Create initial index.json

cat > ~/context-data/personal/index.json << 'EOF'
{
  "version": "2.0",
  "nodes": [],
  "links": [],
  "metadata": {}
}
EOF

3. Start the server

docker compose up -d --build

The server will be available at http://127.0.0.1:3031 (localhost only).

4. Verify health

curl http://localhost:3031/health

Client Integration

OpenCode

Add to ~/.config/opencode/opencode.json:

"mcp": {
  "personal-context": {
    "type": "remote",
    "url": "http://localhost:3031/mcp",
    "enabled": true
  }
}

Claude Desktop

Add to ~/Library/Application Support/Claude/claude_desktop_config.json:

{
  "mcpServers": {
    "personal-context": {
      "command": "curl",
      "args": ["-s", "http://localhost:3031/mcp"]
    }
  }
}

Antigravity AGY

Add to ~/.gemini/antigravity/mcp_config.json:

{
  "mcpServers": {
    "personal-context": {
      "serverUrl": "http://localhost:3031/mcp"
    }
  }
}

Data Structure

Your context data lives in ~/context-data/ on your host machine (bind-mounted into the container). Files in this directory are not included in this repo and should never be committed (already in .gitignore).

~/context-data/
└── personal/
    ├── index.json       # Knowledge graph (nodes + links)
    ├── personal.md      # Your identity/profile
    ├── devices.md       # Hardware info
    ├── servers.md       # Server/infrastructure info
    └── .backup/         # Auto-generated backups

Local Development

npm install
npm run dev       # Hot-reload via tsx
npm run build     # Compile TypeScript
npm run test      # Run Vitest tests

Security Notes

⚠️ This server has NO authentication in its current MVP state. It is designed to run locally only (127.0.0.1:3031). Do NOT expose this port to a public network without adding Bearer token auth first.

  • ~/context-data is in .gitignore — never commit it
  • Port is bound to 127.0.0.1 only in docker-compose.yml
  • Before deploying publicly (e.g., CapRover): add Bearer token auth + HTTPS

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
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
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
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
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
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
e2b-mcp-server

e2b-mcp-server

使用 MCP 通过 e2b 运行代码。

官方
精选
Neon MCP Server

Neon MCP Server

用于与 Neon 管理 API 和数据库交互的 MCP 服务器

官方
精选
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