As part of the keynote presentation on Day 1 of Laracon US, Taylor Otwell introduced Ashley Hindle, who has recently joined the Laravel team and is spearheading Laravel's AI initiatives. Ashley mentioned that they have been speaking with developers to see how they can help make AI better assist you in your day-to-day work with Laravel, without it taking over that work.
Ashley then introduced us to the first such initiative, Laravel Boost, a package you install via Composer, which enhances the pairing of AI with Laravel by helping to keep you in the flow and guiding the AI to work more effectively.
Main Features
Laravel Boost has three core features:
- Provides a Laravel-specific MCP Server: 15 tools to help you in your day-to-day work. Your AI agent can query the database, run code through Tinker or search through documentation
- Version-specific documentation: All the Laravel ecosystem docs have been ingested and vectorised, and so should have fewer hallucinations. This includes things like Inertia, Livewire, Flux, Filament, etc. So you get the most relevant information.
- Laravel maintained AI guidelines: It can create your Cursor rules, Junie guidelines, Github Co-Pilot instructions and your Claude.md files. The team has manually curated those guidelines, including version-specific ones, for Laravel. e.g. guidelines for Inertia 1 or 2 to help the AI agent to behave appropriately in your projects.
Ashley demonstrated these features using the ping-crm demo app. The installation was simple thanks to the interactive installer. It asks what to install, e.g., Boost MCP server, package AI guidelines, Laravel Style AI Guidelines, and Herd MCP server. It also asks which IDE you use. E.g. Claude Code, Cursor, PHP Storm Junie or VS Code. If you already have any of these, it can automatically detect them. And lastly, it asks which AI guidelines to use for your project.
During the demonstration, Ashley asked the AI to "create a new contact using the default contact factory, for a new organisation called 'Laracon is Cool'". The AI agent reviewed the necessary files to understand the models and database schema. It also queried the database to check if there were any existing accounts and used the Tinker tool to utilise the contact and organisation factories to create what we requested.
Interestingly, a white screen of death appeared, and so Ashley asked the AI agent to help him resolve the issue. It retrieved the last error from the logs, read the previous few entries, and also checked the streamed web browser logs using the browser-logs tool. The error was a typo in the app.js file.
The next part of the demo highlighted the benefits of Boost, providing access to the latest documentation from the Laravel ecosystem. The example used was implementing a Deferred component in VueJS with Inertia. While the AI could provide a general solution, it wasn't quite the "Inertia way", but with the use of the search-docs MCP tool, it was able to implement the solution better by specifically following version 2.0 of the Inertia documentation.
The Laravel team has promised that Boost will always be updated with the latest docs for the Laravel ecosystem. You can also chat with the docs if you’d like to.
The final demo involved asking the AI to implement per-second rate limiting to the app. It once again used the search-docs tool and was able to add the feature to the AppServiceProvider, but thanks to the Laravel Boost guidelines, the AI agent went ahead and also wrote tests for that new feature. Without Boost, Cursor may not have thought to write a test.
The rules provided by Boost also contain core rules that apply across every project, regardless of the tech stack you are using. For example, it knows to use the Laravel conventions.
The team built an MCP server package, and version 1.0 will be released soon. This package enables you to build your own MCP server and integrate it into your app by adding a single line in routes/ai.php—a big thanks to Peter Suhm for contributing to that.
Without a doubt, this will be a big win for Laravel Devs as Boost makes coding with AI in Laravel easy to set up and makes the AI agent output much higher quality. Laravel Boost will be open-source and released for free within the next two weeks. You can also contribute to it because we will host the code on GitHub.
We look forward to the official release so we can test it out.
