Ears everywhere
in your stack.

AI agents are blind without context

When a developer asks an AI agent "why is my login broken?", the agent has no access to the browser console, network tab, or server terminal. The developer becomes a copy-paste bridge.

# What the AI agent sees without Fennec:
"I can't access your terminal"
"Please paste the error message"
"I don't have browser access"

# What the developer sees:
✗ TypeError: jwt.sign is undefined          // Browser Console
✗ Error: JWT_SECRET not set in env          // Terminal
✗ POST /api/login -> 500                 // Network Tab

Installation & Configuration

Get Fennec running in under 2 minutes. No external services, no cloud accounts — everything runs locally.

Quick Install

# Install globally
npm install -g @plumpslabs/fennec-cli

# Install browser engines (optional)
fennec install-browsers

# Start the MCP server
fennec start

MCP Client Config

Add to your MCP client's config file. Connect via stdio (default) or SSE (--sse).

{
  "mcpServers": {
    "fennec": {
      "command": "fennec",
      "args": ["start"]
    }
  }
}

{
  "mcpServers": {
    "fennec": {
      "command": "fennec",
      "args": ["start", "--sse"]
    }
  }
}

Client compatibility:

💡 Run fennec start --sse — Fennec outputs the exact remote config snippet on startup. For OpenCode, you can add it as a local stdio server or add type: remote, url: http://localhost:3333/sse to ~/.config/opencode/opencode.json for SSE.

Quick Start

Once Fennec is running and connected to your AI agent, pipe your server output and let the AI investigate.

# Terminal: Start your app with Fennec watching
npm run dev 2>&1 | fennec pipe --name "my-app"

Then ask your AI agent: "Why is my app broken?"

The AI will automatically use Fennec tools to open the browser, check console errors, inspect network requests, correlate with server logs, and report the root cause.

Core Features

165+ MCP tools across 18 categories — browser automation, process management, multi-language debugging, and mobile device control.

AI-Native Debugging for 10+ Runtimes

Fennec's Debug Engine provides three levels of debugging — from passive log analysis to active breakpoint debugging to proactive auto-debug — across JavaScript, Python, PHP, Java/Kotlin, Go, .NET, Ruby, Rust, and more.

Multi-Language Support

Record & Replay (Cassette)

Record MCP tool call sessions and replay them to detect regressions:

# Record a debugging session
debug_record_session("login-bug")
# ... perform debug steps ...
debug_stop_recording()

# Replay and compare
debug_diff_sessions(cassette_a, cassette_b)

Full-Stack Correlation

The crown jewel of Fennec. A single command correlates browser errors with server logs and process state to identify root causes automatically.

// diagnose_fullstack({ processId: "dev-server" })
{
  "correlation": {
    "rootCause": "JWT_SECRET environment variable missing",
    "confidence": 0.94,
    "fix": "Add JWT_SECRET to your .env file",
    "timeline": [
      { "t": "+0ms",  "layer": "browser", "event": "POST /api/auth/login initiated" },
      { "t": "+12ms", "layer": "server",  "event": "ERROR: JWT_SECRET not set" },
      { "t": "+16ms", "layer": "browser", "event": "TypeError thrown in auth.js:67" }
    ]
  }
}

Tool Reference

165+ tools organized into 18 categories. Every tool follows MCP conventions with structured Zod schemas and token-efficient responses.

See the full tool reference for detailed schemas and examples.