Grok PHP is a robust, flexible, and feature-rich PHP library designed to interact seamlessly with the Grok AI API.
- Seamless Integration: Elegant PHP-first interface with intuitive methods for Grok AI
- Advanced Chat Capabilities: Full support for multi-turn conversations and real-time streaming
- Comprehensive Features: Text completions, image analysis, and embeddings all in one!
- Robust Architecture: Type-safe implementations with extensive error handling
- Model Flexibility: Support for multiple Grok models (Grok-2, Grok-2-Vision, etc.)
- Enterprise Ready: Secure API handling with proper authentication
- Response Management: Rich response objects with detailed analytics and metadata
- Easy Configuration: Simple setup with minimal dependencies
- PHP 8.1 or higher
- Composer
- Grok AI API key
Install Grok PHP via Composer:
composer require alvincoded/grok-php-clientHere's how simple it is to use Grok PHP :
<?php
require_once 'vendor/autoload.php';
use GrokPHP\Client\GrokClient;
use GrokPHP\Params;
$client = new GrokClient();
// Simple chat
$response = $client->chat()->generate("Tell me a joke about AI");
echo $response->getContent();
// With system message
$response = $client->chat()->generate(
"What's the best programming language?",
Params::create()
->systemMessage('You are an experienced programmer.')
->temperature(0.7)
);
// Streaming response
$client->chat()->streamChat(
'Tell me something about Grok PHP',
function (ChatMessage $chunk) {
echo $chunk->text();
}
);
// Multi-turn conversation
$chat = $client->beginConvo();
$response = $chat->send('What is machine learning?');
echo $response->text();
$response = $chat->send('Give me an example');
echo $response->text();<?php
require_once 'vendor/autoload.php';
use GrokPHP\Client\GrokClient;
use GrokPHP\Params;
$client = new GrokClient();
// Basic completion
$response = $client->completions()->create(
"The future of AI will",
Params::create()->maxTokens(100)->temperature(0.7)
);
// Multiple completions
$responses = $client->completions()->createMultiple(
"Write a creative title for a sci-fi novel",
3,
Params::create()->temperature(1.0)
);
// Get token count
$tokenCount = $client->completions()->getTokenCount("Sample text");<?php
require_once 'vendor/autoload.php';
use GrokPHP\Client\GrokClient;
use GrokPHP\Params;
$client = new GrokClient();
// Basic image analysis
$response = $client->images()->analyze('https://picsum.photos/200/300');
// Detailed analysis with prompt
$response = $client->images()->analyze(
'https://picsum.photos/200/300',
'What objects can you identify in this image?',
Params::create()->maxTokens(300)->temperature(0.8)
);
// Check image content
$containsPeople = $response->containsContent('person');<?php
require_once 'vendor/autoload.php';
use GrokPHP\Client\GrokClient;
$client = new GrokClient();
$embeddingResponse = $client->embeddings()->create('Hello, world!');
$embeddings = $embeddingResponse->getEmbeddings();<?php
require_once 'vendor/autoload.php';
use GrokPHP\Client\GrokClient;
use GrokPHP\Enums\Model;
$client = new GrokClient();
// Simple chat (with model specification)
$response = $client->model(Model::GROK_2_1212)->generate('Tell me a joke about AI');
echo $response->text();
// Get model capabilities
$model = Model::GROK_2_1212
$config = $client->getConfig();
echo $config->getModelMaxTokens($model) // 32,768
echo $config->modelSupportsStreaming($model) // true
echo $config->modelSupportsFunctions($model) // false<?php
use GrokPHP\Client\GrokClient;
use GrokPHP\Enums\Model;
// Scenario example: A university library needs to process 50,000 research papers into their new digital repository.
// Each entry requires consistent metadata fields.
// 1. Define schema once
$jsonSchema = [
"type" => "object",
"properties" => [
"title" => ["type" => "string"],
"authors" => ["type" => "array", "items" => ["type" => "string"]],
"publication_year" => ["type" => "integer"],
"doi" => ["type" => "string"],
"keywords" => ["type" => "array", "items" => ["type" => "string"]],
"citation_count" => ["type" => "integer"]
],
"required" => ["title", "authors"]
];
// 2. Process documents
$client = new GrokClient();
foreach ($researchPapers as $paperText) {
$metadata = $client->chat()->generateStructured($paperText, $jsonSchema);
// 3. Directly store structured data
$this->database->insertPaper(
title: $metadata['title'],
authors: $metadata['authors'],
year: $metadata['publication_year'] ?? null,
doi: $metadata['doi'] ?? '',
keywords: $metadata['keywords'] ?? []
);
}// Define your schema as a PHP class
class ResearchPaper extends \GrokPHP\Utils\DataModel
{
#[SchemaProperty(type: 'string', description: 'Paper title')]
public string $title;
#[SchemaProperty(type: 'array', description: 'List of authors')]
public array $authors;
#[SchemaProperty(type: 'integer', description: 'Year of publication', required: false)]
public int $publicationYear;
}
// ...then, in your application code
$result = $client->chat()->generateStructured(
"Extract research paper details",
ResearchPaper::class
);
// ...and finally, get typed properties
echo $result->title;
echo $result->authors[0];$response->getContent(); // Get response content
$response->getRole(); // Get message role
$response->getFinishReason(); // Get completion finish reason
$response->getId(); // Get response ID
$response->getModel(); // Get model used
$response->getUsage(); // Get token usage statistics$response->getAnalysis(); // Get analysis text
$response->getImageUrl(); // Get analyzed image URL
$response->getMetadata(); // Get image metadata
$response->getUsage(); // Get token usage$response->getEmbeddings(); // Get embeddings
$response->getUsage(); // Get token usageuse GrokPHP\Exceptions\GrokException;
try {
$response = $client->chat()->generate("Your prompt");
} catch (GrokException $e) {
echo "Error: " . $e->getMessage();
}| Model | Supports Streaming | Supports Functions |
|---|---|---|
| grok-beta | Yes | Yes |
| grok-2-vision-1212 | No | No |
| grok-2-1212 | Yes | Yes |
temperature(float $value): Sets the temperature for sampling the next token.maxTokens(int $value): Sets the maximum number of tokens to generate in the completion.topP(float $value): Sets the top P value for nucleus sampling.stream(bool $value): Sets the presence of streaming responses.systemMessage(string $message): Sets the system message for the AI model.n(int $value): Sets the number of completions to generate.presencePenalty(float $value): Sets the presence penalty.frequencyPenalty(float $value): Sets the frequency penalty.logitBias(array $values): Sets the logit bias for the completion.stop(array $values): Sets the stop sequence for the completion.logprobs(int $value): Sets the logprobs parameter.dimensions(int $value): Sets the dimensions parameter for embedding.echo(bool $value): Sets the echo parameter.user(string $value): Sets the user parameter.suffix(string $value): Sets the suffix that is appended to the completion.
Add the following to your .env file:
GROK_API_KEY=your-api-key
composer testPlease see CHANGELOG for more information on what has changed recently.
Contributions are highly aappreciated! Please see the Contributing Guide for details.
Please review the security policy on how to report security vulnerabilities.
Grok PHP is an open-sourced software licensed under the MIT license.
If you encounter any issues or have questions, please open an issue on the GitHub repository.
Built with ❤️ for the AI community.