The AI-powered performance engine for modern frontend teams.
Catch layout thrashing, memory leaks, and render explosions before they reach production.
Quick Start • Why FPL? • Features • Architecture • Contributing
Get up and running with the CLI and MCP server in under two minutes:
# 1. Install the CLI globally
npm install -g @smg99/frontend-performance-lab-cli
# 2. Configure your IDE (Cursor, Claude, VS Code, Windsurf)
fpl setup
# 3. Verify installation
fpl doctorRestart your IDE and you're ready to go!
- ✓ Installs the
@smg99/frontend-performance-lab-cliglobally - ✓ Detects your active IDE (Cursor, Claude Desktop, Windsurf, etc.)
- ✓ Configures the MCP server automatically in your IDE settings
- ✓ Verifies the connection and dependencies with
fpl doctor - ✓ Ready to use immediately to analyze your components
Traditional AI assistants understand code. Frontend Performance Lab understands frontend performance.
Modern frameworks like React and Vue are fast by default — until they aren't. A single deep array watcher or a forced synchronous layout can crash your Lighthouse score and ruin the user experience.
Frontend Performance Lab provides your AI assistant (via MCP) with a massive, specialized knowledge graph dedicated entirely to making browsers render frames faster and consume less memory.
- Vue, React, Nuxt, Next.js, Angular
- Cursor, Claude Desktop, VS Code, Windsurf
Drop in any .vue, .jsx, or .js file. The engine uses Babel AST traversal to instantly detect layout thrashing, event listener leaks, massive render trees, and reactivity pitfalls.
The full knowledge base — experiments, recipes, browser APIs — exposed directly to your AI assistant. Ask Cursor or Claude to "optimize this component using Frontend Performance Lab" and watch the magic happen.
Live Vue and React components you can tweak in real time. See what happens to frame rates when you skip virtualization on a 10,000-row list, or trigger a forced synchronous layout.
Structured tutorials mapped directly to the analyzer's findings. Fix what the analyzer catches with step-by-step implementation guides.
Analyzer Workspace
|
|
Install Wizard
|
MCP Playground
|
At its core, Frontend Performance Lab is powered by a unified Node.js engine shared seamlessly between the Nuxt 4 web dashboard and the stdio MCP server.
graph TD
A[Browser / Developer] --> B(Nuxt 4 Frontend)
C[AI Assistant / IDE] --> D(MCP Server — stdio)
B --> E{Shared Core}
D --> E
E --> F[AST Analyzer Engine]
E --> G[Knowledge Graph]
E --> H[Browser API Registry]
F --> I[(Local File System)]
Your AI sees exactly what the dashboard sees.
We welcome contributions! Whether you want to add a new performance recipe, write an AST rule, or improve the UI, we'd love your help.
Read our Contributing Guide to get started with local development.
Built with ❤️ by performance enthusiasts.
Visit the WebsiteNew npm releases are created automatically by pushing a version tag.
git tag v1.0.0-beta.2
git push origin v1.0.0-beta.2The publish.yml workflow will publish the CLI to npm using npm Trusted Publisher (OIDC) and create a GitHub Release with autogenerated notes.
If this project saved you time or you found it useful, consider buying me a coffee. It helps me keep building and open-sourcing useful tools.


