Welcome to Limelight
Limelight is the runtime context layer for AI-assisted development. It captures everything that happens when your app runs — network requests, state changes, re-renders, console logs, timing, and errors — and makes it machine-queryable through structured, pre-analyzed context called Debug IR. AI coding assistants can read your source code, but they can’t see what actually happens at runtime. Limelight bridges that gap.Why Limelight?
- Runtime Context for AI — Your AI assistant debugs with real runtime data, not guesses. Ask “why is my search showing stale results?” and get an answer backed by actual causal chains.
- Zero Config Setup — One line of code. No rebuilds, no heavy instrumentation.
- Full-Stack Correlation — Automatically links frontend requests to backend responses to state updates to re-renders across system boundaries.
- Causal Analysis — Limelight doesn’t just capture events. It correlates them into causal chains, detects anti-patterns (N+1 queries, render loops, race conditions), and delivers pre-analyzed context.
- Privacy-First — All data stays on your machine. No telemetry, no cloud dependency.
- Two Surfaces, One Engine — Use the Desktop App for visual debugging or the MCP Server to pipe runtime context directly into Cursor, Claude Code, or any MCP-compatible editor.
Key Features
MCP Server
Pipe live runtime context into your AI coding assistant. Your AI calls
Limelight’s tools automatically to diagnose issues with real data.
Full-Stack Tracing
Automatically correlate client requests with server handling. See the full
request lifecycle from browser to server and back.
Render Tracking
Component renders are captured, scored, and analyzed. Detect render loops,
unnecessary re-renders, and unstable props automatically.
Network Inspection
Every request and response with timing, status, headers, bodies, and GraphQL
intelligence including complexity scoring.
State Debugging
Zustand and Redux store contents, diffs, and change history. See exactly how
state evolved and what triggered each change.
Issue Detection
Automatic detection of N+1 queries, render cascades, race conditions, retry
storms, stale closures, and other runtime anti-patterns.
Two Ways to Use Limelight
MCP Server — AI-Powered Debugging
Give your AI coding assistant direct access to your app’s runtime. The MCP server streams live data from your running app and exposes 11 debugging tools that your AI calls automatically. Works with Cursor, Claude Code, Windsurf, and any MCP-compatible editor.Desktop App — Visual Debugging
Free, local-first visual debugger with an interactive timeline, real-time updates, and built-in AI analysis.Supported Platforms
| Platform | Status |
|---|---|
| React Native | Supported |
| React (web) | Supported |
| Next.js | Supported |
| Node.js / Express | Supported |
| Expo | Supported |