MCP Observability Platform

Overview

Open, vendor-agnostic observability and trust layer for Model Context Protocol (MCP) LLM applications. No vendor lock-in, plug-and-play SDKs, and extensible for rate limiting, tool filtering, auth, and traceability.

🚀 Production Ready: Available at https://etalesystems.com

Features

• Trace collection SDKs (Python, TypeScript)
• Node.js/Express backend with Postgres (Prisma)
• React frontend for trace inspection, replay, and filtering
• Extensible schema for future features (auth, rate limiting, tool filtering)

Quickstart

Option 1: Use Production (Recommended)

1. Get API Key: Visit https://etalesystems.com and create an API key
2. Install SDK:

   # Python
   pip install mcp-hack
   
   # TypeScript/JavaScript
   npm install @mcp-hack/typescript

3. Use in your code: See Python SDK or TypeScript SDK documentation

Option 2: Local Development

1. Backend:

   • cd backend && npm install
   • Set up Postgres and .env (see .env.example)
   • npx prisma migrate dev
   • npm run dev

2. Frontend:

   • cd frontend && npm install
   • npm start

3. SDK Usage: Point to http://localhost:3001 instead of production URL

---

Architecture

• SDKs: Log traces from your MCP/LLM app
• Backend: Store and serve traces
• Frontend: Inspect, replay, and filter traces

---

Extending

• Add fields to the trace schema for rate limiting, tool filtering, auth, etc.
• Add new endpoints or UI features as needed.

---

Examples

See the Hello World MCP Server example that demonstrates:

• Basic MCP server with observability
• Math tools (add, multiply, divide)
• Real-time trace collection
• Dashboard integration

Links

• 🌐 Dashboard: https://etalesystems.com
• 📚 Python SDK: PyPI
• 📚 TypeScript SDK: NPM
• 🐙 Source Code: GitHub