Weave Code
Boost Laravel Package
laravel/boost
Laravel Boost accelerates AI-assisted Laravel development by supplying the context, conventions, and structure AI needs to generate high-quality, framework-specific code and suggestions, helping teams build faster while staying aligned with Laravel best practices.
Laravel-focused MCP server for augmenting your AI powered local development experience.
Frequently asked questions about Boost
- How do I install Laravel Boost in an existing Laravel project?
- Run `composer require laravel/boost` followed by `php artisan boost:install`. This sets up the default configuration and installs core skills. If you’re using a custom project structure, you may need to adjust the `boost.json` paths manually after installation.
- Which Laravel versions does Boost support, and what if I’m on Laravel 9?
- Boost officially supports Laravel 10.x and above, with some features requiring Laravel 11+ or 13+. If you’re on Laravel 9, you’ll need to avoid skills or configurations that rely on newer Laravel conventions, like Livewire v4 or MCP integrations.
- Can I use Boost with GitHub Copilot or other AI tools besides Kiro IDE?
- Yes, Boost is agent-agnostic and works with any AI tool that supports Model Context Providers (MCP). GitHub Copilot, Claude Code, and other MCP-compatible agents can integrate via the `mcp_config_path` in `boost.json`. Check your agent’s documentation for MCP setup instructions.
- What are ‘skills’ in Laravel Boost, and how do I add custom ones?
- Skills are modular packages that define Laravel-specific best practices, testing guidelines, or framework features (e.g., `laravel-best-practices`, `pest-testing`). To add custom skills, place them in `resources/boost/skills`, then run `php artisan boost:update --discover` to register them.
- Will Boost slow down my Laravel application in production?
- No, Boost has minimal runtime overhead. Skills are lazy-loaded, and the token-based class parsing only runs during code generation tasks (e.g., AI-assisted development). Production performance remains unaffected unless you’re actively using AI tools integrated with Boost.
- How do I enforce testing requirements for AI-generated code in Boost?
- Enable the `enforce_tests` flag in `boost.json` to require test files for generated code. You can also create a custom skill (e.g., `pest-testing`) that enforces specific testing patterns. Boost will validate compliance during code generation.
- Can I use Laravel Boost in a monorepo with multiple Laravel projects?
- Yes, Boost supports monorepos by allowing you to specify a custom `mcp_config_path` in `boost.json`. This isolates MCP configurations per project. Run `php artisan boost:validate` to ensure your setup is correct across all projects.
- What happens if my AI agent (e.g., GitHub Copilot) fails or returns errors?
- Boost includes retry logic for agent failures, but you may need to handle rate limits or API errors manually. For critical workflows, consider falling back to a local MCP or implementing custom error handling in your `boost.json` configuration.
- Are there alternatives to Laravel Boost for AI-assisted Laravel development?
- Alternatives include generic AI tools like GitHub Copilot (without Boost’s Laravel context) or custom scripts using Laravel’s token parsing. However, Boost is the only Laravel-specific solution designed to provide framework-aware context, skills, and seamless integration with MCP agents.
- How do I test AI-generated code in CI/CD pipelines with Boost?
- Use `php artisan boost:update --discover` in your CI pipeline to validate skills and configurations. For testing generated code, integrate Pest or PHPUnit tests triggered by Boost’s `enforce_tests` flag. You can also use Boost’s `boost:validate` command to catch misconfigurations early.
Weaver
How can I help you explore Laravel packages today?