limelight-mcp
v1.3.3
Published
Limelight MCP Server — runtime debugging intelligence for AI coding assistants
Downloads
1,814
Maintainers
Readme
limelight-mcp
Runtime debugging context for AI coding assistants. Limelight captures renders, state changes, network requests, and console logs from your React/React Native app and pipes them into Cursor, Claude Code, or any MCP-compatible editor.
Documentation
📚 Full documentation at docs.getlimelight.io
Setup
1. Add the MCP server to your editor
Claude Code:
claude mcp add limelight-mcp npx limelight-mcpCursor / other MCP clients:
{
"mcpServers": {
"limelight": {
"command": "npx",
"args": ["limelight-mcp"]
}
}
}2. Install the SDK in your app
npm install @getlimelight/sdkimport { Limelight } from "@getlimelight/sdk";
Limelight.connect({
target: "mcp",
});3. Run your app and ask your AI
"My app feels slow. Do you see any issues?"
Limelight captures runtime data from your running app and makes it available to your AI assistant — renders, state updates, network requests, and console logs with full correlation and analysis.
Tools
| Tool | Description |
| --------------------------------- | ---------------------------------------------- |
| limelight_get_session_overview | High-level snapshot of your running app |
| limelight_find_issues | Proactive scan for performance issues and bugs |
| limelight_investigate_error | Root cause analysis with causal chains |
| limelight_query_network | Query captured network requests |
| limelight_query_logs | Query console events |
| limelight_get_render_profile | Component render performance profiling |
| limelight_get_state_snapshot | State store contents and history |
| limelight_investigate_component | Deep-dive on a specific component |
| limelight_correlate_event | Find related events by correlation |
| limelight_get_timeline | Chronological event view |
Telemetry
Limelight collects anonymous usage data to help us improve the MCP server. Telemetry is enabled by default and only sends:
- Which MCP tools are called (tool names only)
- How long each tool call takes (duration in ms)
- Server start events
No query content, runtime data, Debug IR content, or any data from your app is ever collected.
An anonymous device ID is generated and stored at ~/.limelight/device_id. No personally identifiable information is sent.
To opt out, set the environment variable before starting the server:
LIMELIGHT_TELEMETRY=falseFor example, in your MCP client config:
{
"mcpServers": {
"limelight": {
"command": "npx",
"args": ["limelight-mcp"],
"env": {
"LIMELIGHT_TELEMETRY": "false"
}
}
}
}Requirements
- Node.js 18+
- An MCP-compatible AI client (Claude Code, Cursor, etc.)
