Google's Gemini AI In Laravel

Installation

To use Google's Gemini API, you must have an API key. To obtain an API key, go to Google AI Studio.

Installing the Gemini API Client for Laravel with Composer is the first step.

composer require gemini-api-php/laravel

Configuration

The client can be configured in two different ways.

Variables in the environment

You can use the API key you received from Google AI studio to set the GEMINI_API_KEY environment variable.

In your.env file, add the following line.

Configuration file

You can also run the following command to create a configuration file in your applications config folder.

The Gemini API client can now be configured by editing the config/gemini.php file.

Text Generation

    use GeminiAPI\Laravel\Facades\Gemini;

    print Gemini::generateText('PHP in less than 100 chars');            
            
    // PHP: A scripting language that runs on servers that is used to make dynamic websites.
    // Open-source, often used, and simple to learn.                
                

Text Generation Using Image File

    use GeminiAPI\Laravel\Facades\Gemini;

    print Gemini::generateTextUsingImageFile(
        'image/jpeg',
        'horse.jpg',
        'Explain what is in the image',
    );
    
    // A horse is seen in the picture standing on Earth.
    // The blazing emblem on the forehead of the metal horse.
    // There's a web of luminous lines all around the Earth.
    // The background of the image is covered in stars.
        

Text Generation Using Image Data

    use GeminiAPI\Laravel\Facades\Gemini;

    print Gemini::generateTextUsingImage(
        'image/jpeg',
        base64_encode(file_get_contents('horse.jpg')),
        'Explain what is in the image',
    );            
    
    // A horse is seen in the picture standing on Earth.
    // The blazing emblem on the forehead of the metal horse.
    // There's a web of luminous lines all around the Earth.
    // The background of the image is covered in stars.
    
        

Chat Session (Multi-Turn Conversations)

    use GeminiAPI\Laravel\Facades\Gemini;

    $chat = Gemini::startChat();
    
    print $chat->sendMessage('Hello Techvoot in PHP');                   
    
    // echo "Hello Techvoot!";
    // This code will print "Hello Techvoot!" to the standard output.        
    
    print $chat->sendMessage('in Go');
    
    // fmt.Println("Hello Techvoot!")
    // This code will print "Hello Techvoot!!" to the standard output.
    

Chat Session with History

    use GeminiAPI\Laravel\Facades\Gemini;

    $history = [
        [
            'message' => 'Hello Techvoot in PHP',
            'role' => 'user',
        ],
        [
            'message' => << 'model',
        ],
    ];
    $chat = Gemini::startChat($history);
    
    print $chat->sendMessage('in Go');                              
    
    // fmt.Println("Hello Techvoot!")
    // This code will print "HelloTechvoot!" to the standard output.
        

Text Embeddings

    use GeminiAPI\Laravel\Facades\Gemini;

    print_r(Gemini::embedText('PHP in less than 100 chars'));                                        
    
    // [
    //    [0] => 0.041395925
    //    [1] => -0.017692696
    //    ...
    // ]    
        

Tokens counting

    use GeminiAPI\Laravel\Facades\Gemini;
    print Gemini::countTokens('PHP in less than 100 chars');                                       
    
    // 10      
        

Listing models

    use GeminiAPI\Laravel\Facades\Gemini;

    print_r(Gemini::listModels());                                          
    
    //[
    //  [0] => GeminiAPI\Resources\Model Object
    //    (
    //      [name] => models/gemini-pro
    //      [displayName] => Gemini Pro
    //      [description] => The best model for scaling across a wide range of tasks
    //      ...
    //    )
    //  [1] => GeminiAPI\Resources\Model Object
    //    (
    //      [name] => models/gemini-pro-vision
    //      [displayName] => Gemini Pro Vision
    //      [description] => The best image understanding model to handle a broad range of applications
    //      ...
    //    )
    //]          
        

Integrating Gemini Pro Vision in Laravel

Thanks to the user-friendly interface offered by the Gemini PHP package, using Gemini Pro Vision for image or video analysis is simple. You can ask the AI to interpret an image in the following ways:

    use Gemini\Laravel\Facades\Gemini;
    use Gemini\Types\Blob;
    use Gemini\Types\MimeType;
    
    $result = Gemini::geminiProVision()->generateContent([
        'What is this picture?',
        new Blob(
            mimeType: MimeType::IMAGE_JPEG,
            data: base64_encode(file_get_contents('your-image-url.jpg'))
        )
    ]);
    
    echo $result->text();            
        

Here, we are requesting that Gemini Pro Vision examine a picture and provide a description of what it observes. In addition to the picture data you submit as a Blob object with the base64-encoded image, the API expects a question. This capability can completely change the way users engage with your application by enabling features such as content-based recommendation systems, automatic image tagging, and even accessibility improvements through the provision of text descriptions for visual material.