MemScope

Real-time memory profiler for backend & frontend applications

MemScope helps developers visualize, understand, and debug memory usage in their projects with zero setup and a live dashboard.

It works locally, runs entirely on your machine, and supports:

• Backend (Node.js)
• Frontend (Browser JS heap – Chromium-based)
• Full‑stack apps (backend + frontend together)

Why MemScope?

Memory bugs are hard:

• Silent leaks
• Unpredictable spikes
• Heap snapshots that are painful to interpret
• No unified view for backend + frontend

MemScope fixes that by giving you:

• Live memory graphs
• Clear spikes & GC behavior
• Backend + frontend separation
• One-command usage
• No cloud, no accounts, no config

Installation & Usage

Option 1: One-time run (recommended)

npx memscope run node app.js

Option 2: Install globally

npm install -g memscope
memscope run node app.js

This automatically:

• Starts the MemScope collector
• Injects the memory agent
• Opens a live dashboard
• Streams memory data in real time

Tracking Modes

Backend only (default)

memscope run node app.js

Tracks:

• Node.js heap
• RSS / external memory
• GC behavior

Frontend only (browser)

memscope run --frontend

Then inject the browser agent in development only.

React / Vite

if (import.meta.env.DEV) {
  import("memscope/agent/browser");
}

CRA / Webpack

if (process.env.NODE_ENV === "development") {
  require("memscope/agent/browser");
}

⚠️ Frontend tracking works best in Chrome / Edge. Safari does not expose JavaScript heap APIs.

Full‑stack (backend + frontend)

memscope run --both npm run dev

This gives you two live graphs:

• Backend memory
• Frontend browser memory

Perfect for:

• SSR leaks
• API + UI interactions
• Long‑running sessions

Dashboard

Once running, open:

http://localhost:3333/dashboard/index.html

You’ll see:

• Live memory timeline
• Backend vs frontend separation
• GC dips & spikes
• Clear growth patterns

Example Use Cases

Debug a memory leak

memscope run node worker.js

Compare before/after refactor

memscope run node app.js
# refactor code
memscope run node app.js

SPA navigation leak

memscope run --frontend
npm run dev

How MemScope Works (Internals)

CLI
 ├─ starts collector (local)
 ├─ injects memory agent
 ├─ runs your app
 │
 ├─ agent samples memory
 ├─ collector aggregates data
 ├─ WebSocket streams live points
 └─ dashboard renders graphs

Everything runs locally.

Privacy & Security

• No cloud
• No uploads
• No tracking
• No external services

MemScope never sends your data anywhere.

Current Limitations

• Frontend heap tracking requires Chromium
• Port 3333 must be free
• Node.js >= 16

Roadmap

• Heap / RSS / External toggles
• Memory spike annotations
• Session comparison
• Leak heuristics
• Python & JVM agents

Author

Built with ❤️ for developers who care about performance.

Contributions welcome.