Installation

Requirements

  • PHP 8.2 or higher. Laravel 13 itself needs PHP 8.3.
  • Laravel 11, 12 or 13.
  • An API key of at least one provider: Claude (Anthropic), ChatGPT (OpenAI), Gemini (Google) or Grok (xAI). Don’t have one yet? Get your API keys explains where, step by step. The provider bills every call to your account; the package itself is free.

The package has no migrations, routes or views, so there is nothing to migrate or to link.

Step 1: install the package

Run this in the root folder of your Laravel application:

composer require darvis/laravel-ai-generator

Laravel discovers the service provider by itself. You don’t register anything.

Step 2: run the setup wizard 

php artisan ai-generator:install

The wizard asks you a few questions and writes the answers to your .env file. It asks:

  1. Which providers you have a key for. Use the arrow keys and the space bar to select one or more, then press Enter.
  2. The API key of each provider. Paste it; you won’t see it on the screen. When a key is already set, the wizard asks whether to keep it.
  3. The wizard then checks the key with the provider. It only asks for the list of models, which costs nothing. When the provider refuses the key, you can enter another one, keep it anyway or skip that provider.
  4. Which model writes the text, and for OpenAI, Gemini and Grok which model makes the images. You choose from the models your key may use. Not sure? Keep the one marked “(current)”.
  5. With more than one provider: which one writes by default, which one makes the images, and which ones take over when the first one fails. See Multiple providers.
  6. Optionally the default language, tone and reading level.
  7. A summary, with the keys partly hidden, and whether to write it to .env.

.env is the file with the settings and secrets of your application. It must never end up in git; Keep your keys secret explains why and how.

Run the wizard again at any time to add a provider or change a model.

Or set the key by hand

Prefer to edit .env yourself? Add the key of your provider, for example:

OPENAI_API_KEY=your-api-key-here

With another provider than OpenAI, also say which driver writes the text:

ANTHROPIC_API_KEY=your-api-key-here
AI_GENERATOR_DRIVER=anthropic

Every other setting has a default; see Configuration.

Step 3 (optional): publish the config file

You only need this when you want to change a setting in PHP instead of in .env.

php artisan vendor:publish --tag=ai-generator-config

This copies the file to config/ai-generator.php in your application.

Check that it works

Do this in two steps. The first one costs nothing.

1. The keys work 

php artisan ai-generator:status

You see a table with the four providers. For every provider with a key, the column Check says OK when the provider accepted the key. This only asks each provider for its list of models, which is free. Below the table you see the default driver, the image driver, the fallbacks and whether the MCP server is available.

  • no key: that provider has no key. That is fine when you don’t use it.
  • ... refused the API key (HTTP 401).: the key is wrong or revoked. See Troubleshooting.
  • key OK, but ... is not in the model list: your key works, but the configured model is not one your key may use. Pick another one with the wizard.

Add --offline to see the settings without contacting any provider.

2. One real call, without an image

This sends one text request to the driver you name, and is billed to your account. It never makes an image.

php artisan ai-generator:status --test=anthropic

Use openai, anthropic, gemini or xai. After some seconds you see the model, the title and the intro it wrote. The words differ on every run.

The same check in PHP, for example in php artisan tinker:

use Darvis\LaravelAiGenerator\ContentRequest;
use Darvis\LaravelAiGenerator\Facades\AiGenerator;

AiGenerator::generate(new ContentRequest(
    topic: 'Why automated tests matter',
    language: 'en',
    maxWords: 100,
    includeImage: false,
))->title;

includeImage: false keeps it to one call and maxWords: 100 keeps the text short. When you see an exception instead, its message tells you what went wrong. Every message is listed in Troubleshooting.

Laravel Boost

The package ships a Laravel Boost guideline and a skill in resources/boost/. Run php artisan boost:install, or php artisan boost:update --discover in a project that already uses Boost. Your AI assistant then knows the API, the defaults and the pitfalls: the image that is generated by default, the providers, the custom driver binding and how to test without calling a real API.

Next steps

  • Usage: a complete example and every request option.
  • Multiple providers: Claude for the text, OpenAI for the images, and fallbacks.
  • MCP server: let an assistant such as Claude Code generate content through your app.
  • Configuration: change a model, the default language or the timeout.