What is Laravel Boost and Why Every Laravel Dev Needs It

Your AI assistant knows Laravel. It doesn't know your Laravel app. That's the gap Laravel Boost fills.

Laravel Boost is an MCP (Model Context Protocol) server that runs inside your Laravel application and exposes its internals to AI coding agents. Routes, database schema, configuration, logs, artisan commands. Boost gives your AI assistant the full picture of your specific project instead of forcing it to guess or ask you twenty questions.

It ships as a dev dependency. No production footprint, no runtime overhead in deployed code.

Two commands:

composer require laravel/boost --dev
php artisan boost:install

The install command auto-detects your IDE and AI agents, then generates the necessary configuration files like .mcp.json and guideline files. That's it. No config files to hand-edit, no environment variables to set.

Boost gives AI agents access to over 15 specialized tools covering most of what you'd need to understand a Laravel app:

• Application Introspection: PHP and Laravel versions, installed packages, Eloquent models, and configuration values (with sensitive values redacted).
• Routes: Every registered route, including middleware, controller, method, and URI. The agent sees your actual routing structure, not a guess based on file names.
• Database Schema: Tables, columns, types, indexes, foreign keys. The agent can write migrations and queries that match your actual database.
• Database Queries: Read-only queries against your database to understand your actual data, not just the schema definition.
• Logs: Recent application logs so the agent can diagnose issues without you copy-pasting stack traces.
• Error Logs: The last error from your application's log files with full context.
• Tinker: A sandboxed REPL the agent can use to inspect runtime state, test Eloquent queries, or verify assumptions about your data.
• Artisan Commands: The full list of available commands, including custom ones, with their signatures and descriptions.
• Browser Logs: Console logs and errors from the browser for full-stack debugging.
• Documentation Search: Access to over 17,000 pieces of Laravel-specific knowledge, covering the framework, ecosystem packages, and common patterns.

That documentation search deserves its own callout.

Most AI models have general Laravel knowledge baked into their training data, but it's frozen at a cutoff date and mixed in with every other framework they've seen. Boost's documentation API gives the agent access to over 17,000 curated, structured pieces of Laravel knowledge, current, searchable, and contextual.

When your agent needs to know how a specific Laravel feature works, it queries this API instead of relying on potentially outdated training data. The difference shows up fast: fewer hallucinated method signatures, correct default values, and awareness of features introduced in recent Laravel versions.

Without Boost, here's what happens when you ask an AI agent to add a feature to your app:

1. It reads some files.
2. It guesses at your database structure from model files (maybe).
3. It writes code that could work in a Laravel app.
4. You spend time fixing the parts that don't match your actual setup.

With Boost:

1. It queries your routes, schema, and config.
2. It understands your actual database structure, registered middleware, and available commands.
3. It writes code that fits your Laravel app.

The difference is stark when you're working with non-trivial applications. A fresh Laravel install is easy for any AI to reason about. A two-year-old app with 200 routes, custom middleware, and 60 database tables is where Boost earns its keep.

// Without Boost, the agent might write this:
Route::get('/users', [UserController::class, 'index']);

// With Boost, it knows you already have:
// - A UserController with specific middleware
// - An 'admin.' route prefix
// - A policy registered for User
// And generates code that actually fits your existing patterns.

After installing Boost, it registers itself automatically. Run your Claude Code session from your project root and the MCP server is available immediately. You can verify with:

claude mcp list

You should see laravel-boost in the output.

Add the MCP server to your .cursor/mcp.json:

{
  "mcpServers": {
    "laravel-boost": {
      "command": "php",
      "args": ["artisan", "boost:mcp"],
      "cwd": "/path/to/your/laravel/app"
    }
  }
}

Restart Cursor after adding the configuration.

In PhpStorm, go to Settings -> Tools -> AI Assistant -> MCP Servers and add a new server:

• Command: php
• Arguments: artisan boost:mcp
• Working Directory: your project root

PhpStorm will connect to the MCP server on the next AI interaction.

If you use AI tools for Laravel development, yes. It's a dev-only dependency, and the context it provides eliminates an entire category of "the AI didn't know about my app" problems.

The smaller your app, the less you need it. The larger and more complex your app, the more time Boost saves by giving the agent real context instead of forcing it to reconstruct your application's architecture from file reads alone.

Try it on your current project and see the difference in your next AI-assisted coding session.

Thanks for reading, and see you in the next one.