smart-xdebug-mcp
Connects AI assistants like Claude to PHP's XDebug debugger for runtime inspection, breakpoint control, and variable tracking, enabling natural language debugging of PHP applications.
README
Smart XDebug MCP Server
Let AI debug your PHP code.
Smart XDebug MCP connects Claude to your PHP application's debugger, enabling AI-assisted debugging that actually understands your code at runtime.
<!--
TODO: Add demo GIF here

To create the demo GIF:
- Install terminalizer: npm install -g terminalizer
- Record: terminalizer record demo
- Edit demo.yml to customize
- Generate: terminalizer render demo -o docs/demo.gif -->
The Problem
Debugging PHP applications traditionally requires:
- Manually setting breakpoints in an IDE
- Stepping through code line by line
- Remembering to check the right variables at the right time
- Context-switching between your AI assistant and your debugger
When you ask an AI to help debug, it can only see your code statically. It can't see what's actually happening when your code runs—what values variables hold, where execution flows, or why that exception was thrown.
The Solution
Smart XDebug MCP gives Claude direct access to your PHP debugger. Instead of guessing what might be wrong, Claude can:
- Set breakpoints at suspicious locations
- Run your code and pause at those breakpoints
- Inspect variables to see actual runtime values
- Step through execution to understand the flow
- Track variable changes over time with built-in history
It's like pair programming with a debugger expert who never gets tired and can analyze complex state instantly.
Why This Matters
For Developers
- Faster debugging — Describe the bug, let AI investigate
- Better insights — AI can correlate patterns across variables you might miss
- Learning tool — Watch how an experienced debugger approaches problems
For AI
- Runtime context — See actual values, not just code structure
- Surgical precision — Request only the data needed, preventing context overload
- Time-travel — Query what variables were at previous breakpoints
How It Works
You: "My /api/orders endpoint returns empty for user 42, but they have orders in the database"
Claude: *Sets breakpoint in OrderController*
*Triggers the API request*
*Inspects $user, $query, $results at the breakpoint*
Claude: "Found it. The query filters by user_id but $user->id is null here
because the auth middleware isn't running on this route."
No manual stepping. No guessing. Just answers.
Quick Start
1. Install
npm install -g smart-xdebug-mcp
2. Configure Claude Code
Add to your MCP settings (~/.config/claude/settings.json or project .mcp.json):
{
"mcpServers": {
"xdebug": {
"command": "npx",
"args": ["smart-xdebug-mcp"]
}
}
}
3. Configure XDebug
Ensure your PHP environment has XDebug configured:
xdebug.mode = debug
xdebug.start_with_request = trigger
xdebug.client_host = host.docker.internal ; For Docker
xdebug.client_port = 9003
4. Debug
Ask Claude to debug your PHP code:
"Set a breakpoint at app/Services/PaymentService.php line 127,
then run: curl 'http://localhost/api/checkout?XDEBUG_SESSION=mcp'"
Features
| Feature | What It Does |
|---|---|
| Natural Language Debugging | Tell Claude what to debug in plain English |
| Smart Variable Inspection | AI requests only relevant data, not full dumps |
| Time-Travel Queries | See what variables were at previous steps |
| Automatic Path Mapping | Works with Docker without manual configuration |
| Exception Detection | Auto-pause when errors occur |
| Session History | Review what happened in past debug sessions |
Use Cases
- "Why is this returning null?" — Trace execution to find where values disappear
- "This loop seems slow" — Set conditional breakpoints to inspect specific iterations
- "The API works locally but fails in Docker" — Debug the containerized app directly
- "I inherited this codebase" — Let AI explore and explain runtime behavior
Requirements
- Node.js 18+
- PHP with XDebug 3.x configured for remote debugging
- Claude Code or another MCP-compatible AI assistant
Configuration
Environment variables for advanced configuration:
| Variable | Default | Description |
|---|---|---|
XDEBUG_MCP_PORT |
9003 |
XDebug listener port |
XDEBUG_MCP_TIMEOUT |
30000 |
Connection timeout (ms) |
XDEBUG_MCP_WATCHDOG_TIMEOUT |
300000 |
Auto-terminate idle sessions (5 min) |
Path mappings are auto-detected from .vscode/launch.json or docker-compose.yml.
Architecture
Smart XDebug MCP acts as a bridge:
Claude ←→ MCP Protocol ←→ Smart XDebug MCP ←→ DBGp Protocol ←→ XDebug ←→ PHP
The server translates high-level debugging intent into DBGp commands, handles the complexity of the protocol, and returns clean, AI-friendly responses.
Contributing
Contributions are welcome! See CONTRIBUTING.md for guidelines.
License
Copyright 2026 Tyler Wall
Licensed under the Apache License, Version 2.0. See LICENSE for details.
Keywords: PHP debugging, AI debugging, Claude MCP, XDebug integration, autonomous debugging, Model Context Protocol, PHP development tools, AI-assisted development, remote debugging, Docker PHP debugging
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.
E2B
Using MCP to run code via e2b.
Neon Database
MCP server for interacting with Neon Management API and databases
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.