glasspulse
Enables AI agents to query and monitor service status, latency, and performance through configurable checks (HTTP, TCP, SSL, keyword, JSON-path) and exposes tools for listing services, retrieving metrics, and running on-demand checks.
README
<div align="center">
◈ GlassPulse
The glassmorphic service dashboard — monitoring so pretty, your status page goes viral.
Checks · Uptime · Latency · Performance — rendered in vivid aurora glass.
</div>
✨ Why GlassPulse?
Status pages look like 1999. GlassPulse is what happens when a design system and a monitoring engine fall in love: frosted-glass cards on an animated aurora gradient, neon status signals, live latency sparklines — and a native MCP server so your AI agent can watch your infrastructure with you.
- 🪟 True glassmorphism — backdrop blur, specular highlights, aurora blobs
- 🎨 Vivid & vibrant — violet → fuchsia → cyan palette, dark-first, light theme included
- 📊 Real monitoring — HTTP, TCP port, keyword, JSON-path and SSL-expiry checks
- ⚡ Performance budgets — services degrade (yellow) when latency exceeds your SLO
- 🧩 Config-driven — one
services.json, JSON-Schema validated, zero backend required - 🤖 MCP-first — ship your dashboard and an agent-ready MCP server from one repo
- 🌗 Theme toggle — dark aurora or light glass, persisted per browser
🚀 Quickstart
git clone https://github.com/MountainLabsDE/GlassPulse-Dashboard.git
cd GlassPulse
npm install
npm run dev # → http://localhost:5173
That's it. The bundled demo config probes real public endpoints (GitHub API, Cloudflare, jsDelivr) so you can see live data immediately.
Configure your services
Edit app/src/config/services.json:
{
"title": "My Infrastructure",
"services": [
{
"id": "api",
"name": "REST API",
"group": "Core",
"icon": "🛰️",
"accent": "#8b5cf6",
"checks": [
{ "type": "http", "url": "https://api.example.com/health", "degradeAboveMs": 800 },
{ "type": "ssl", "host": "api.example.com", "warnDays": 14 }
]
}
]
}
Five check types: http · tcp · keyword · jsonpath · ssl
Full reference: app/schema/glasspulse.schema.json
Browser note: the dashboard runs checks client-side, so CORS-enabled endpoints light up instantly. TCP/SSL checks (and any endpoint behind CORS) are marked agent-only — the MCP server probes them natively via Node.
🤖 MCP Server
GlassPulse ships a first-class MCP server (stdio + streamable HTTP):
npm run mcp:build
node mcp-server/dist/index.js # stdio
GLASSPULSE_MCP_TRANSPORT=http npm run mcp # HTTP :3100/mcp
GLASSPULSE_CONFIG=/path/to/your.json npm run mcp
Tools: list_services · get_status · get_metrics · run_check
Resources: glasspulse://services · glasspulse://status
Claude Desktop / Cursor
{
"mcpServers": {
"glasspulse": {
"command": "node",
"args": ["/absolute/path/to/GlassPulse/mcp-server/dist/index.js"]
}
}
}
Then just ask your agent: "Is the API up? What's the p95 latency?"
📦 Deploy
npm run build # static bundle in app/dist
Drop app/dist on GitHub Pages, Vercel, Netlify or any static host.
For TCP/SSL checks from the dashboard, run the MCP server (or any static
file server) inside your network and point GLASSPULSE_CONFIG at the
same config.
🗺 Roadmap
- [ ] Incident timeline & maintenance windows
- [ ] WebSocket live push (optional tiny agent)
- [ ] Webhook/alert integrations (Slack, Discord, Telegram)
- [ ] Multi-dashboards from one config
- [ ] Plugin checks (custom JS probes)
🤝 Contributing
PRs welcome! Keep the glass aesthetic, add tests for engine logic.
📄 License
Apache-2.0 — © MountainLabs UG (haftungsbeschränkt).
We chose Apache-2.0 over MIT deliberately: it grants contributors and users
an explicit patent license while requiring attribution via NOTICE. The
"GlassPulse" name and logo remain trademarks of MountainLabs UG.
<div align="center"> Made with 💜 by <a href="https://github.com/MountainLabsDE">MountainLabs</a> </div>
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。