sync-vault-agent
An MCP server that enables AI assistants to search, read, and manage files across multiple cloud drives (Baidu, Aliyun, 115, OneDrive, Quark) through a unified interface.
README
<div align="center">
sync-vault-agent
让你的 AI 直接读写百度网盘、阿里云盘、115、夸克、OneDrive
</div>
⚠️ Want to see it in action? Add your screenshot at
docs/hero.png— or try it now with a free account.
What is this?
When your AI assistant (Claude Code, Cursor, Windsurf) needs to find a file on your cloud drive — a quarterly report on Baidu, a meeting note on OneDrive — you want it to just find it. Not switch between 5 browser tabs, not download and re-upload.
sync-vault-agent is the missing link. An MCP server that connects AI clients to your cloud storage accounts. One npm install, one config, and your AI can search, read, and manage files across 5 cloud drives as if they were local.
English · 中文
Demo
"Find the Q4 report in Baidu, search all drives for budget.xlsx, read yesterday's meeting notes from OneDrive"
<!-- TODO: Add a screen recording GIF showing:
- User asks AI to find a file
- AI searches across connected drives
- Returns results with file preview -->
Supported Drives
| Drive | Auth | Status |
|---|---|---|
| 百度网盘 (Baidu) | OAuth | ✅ |
| 阿里云盘 (Aliyun) | OAuth | ✅ |
| 115 网盘 | OAuth | ✅ |
| OneDrive | OAuth | ✅ |
| 夸克网盘 (Quark) | Cookie | ✅ |
Quick Start
# 1. Install the agent globally
npm install -g sync-vault-agent
# 2. Initialize (creates config + opens browser for login)
mcp-agent init
# 3. Add to your AI client's mcp.json
Add to claude_desktop_config.json, ~/.cursor/mcp.json, or VS Code Continue config:
{
"mcpServers": {
"unidrive": {
"command": "mcp-agent",
"args": ["start"]
}
}
}
That's it. Restart your AI client and you can start asking questions about your cloud files.
Why sync-vault-agent?
The problem
- Your AI can only access files on your local machine
- Cloud drive files are fragmented across 5+ different ecosystems
- Switching between drive UIs kills productivity
The solution
sync-vault-agent gives your AI a unified gateway to all your cloud storage:
- Multi-cloud, one agent — connect Baidu, Aliyun, 115, OneDrive, Quark through a single MCP server
- 10+ AI clients supported — Claude Code, Cursor, Windsurf, VS Code Continue, any MCP-compatible tool
- 30-second setup — install, init, done
- Zero dependencies — pure Node.js, no bloated SDKs
- Secure by design — refresh tokens encrypted server-side, access tokens rotated every 2 hours, agent holds short-lived credentials only
- Usage dashboard — real-time analytics, per-drive breakdown
Configuration
Config lives at ~/.config/sync-vault/agent.json. Override via env vars:
| Variable | Default | Description |
|---|---|---|
SYNC_VAULT_API_KEY |
— | API key from Dashboard |
SYNC_VAULT_AUTH_URL |
https://auth.sync-vault.com |
Auth service URL |
SYNC_VAULT_STORE_PATH |
~/.config/sync-vault/ |
Credential store path |
Build from Source
git clone https://github.com/abcamus/sync-vault-agent.git
cd sync-vault-agent
npm install
npm run build
Contributing
PRs welcome! Please open an issue first to discuss changes.
Links
License
MIT © Camus Qiu
推荐服务器
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。
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。
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。
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。
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。
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。
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。
e2b-mcp-server
使用 MCP 通过 e2b 运行代码。
Neon MCP Server
用于与 Neon 管理 API 和数据库交互的 MCP 服务器
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。