Building AI-Powered Apps in PHP — A Step-by-Step Guide

Published 2026-03-18 · Wingman Protocol

This tutorial walks you through creating AI-powered applications using PHP and HTTP clients. We'll cover installation, authentication, and practical examples including chat completions and SEO audits.

Installation and Setup Start by installing Guzzle HTTP client via Composer: composer require guzzlehttp/guzzle

// config.php
define('API_BASE_URL', 'https://api.wingmanprotocol.com/v1');
define('API_KEY', 'your_api_key_here');

function createApiClient() {
    return new \GuzzleHttp\Client([
        'base_uri' => API_BASE_URL,
        'headers' => [
            'Authorization' => 'Bearer ' . API_KEY,
            'Content-Type' => 'application/json',
        ]
    ]);
}

<?php
require_once 'vendor/autoload.php';
require_once 'config.php';

function generateChatResponse($messages) {
    $client = createApiClient();
    
    try {
        $response = $client->post('/chat/completions', [
            'json' => [
                'model' => 'gpt-4',
                'messages' => $messages,
                'max_tokens' => 2000,
            ]
        ]);
        
        return json_decode($response->getBody(), true);
    } catch (\Exception $e) {
        error_log("Chat API error: " . $e->getMessage());
        return null;
    }
}

// Usage
$messages = [
    ['role' => 'system', 'content' => 'You are a helpful assistant.'],
    ['role' => 'user', 'content' => 'Explain quantum computing simply.'],
];

$result = generateChatResponse($messages);
echo $result['choices'][0]['message']['content'];

function seoAudit($url) {
    $client = createApiClient();
    
    try {
        $response = $client->post('/seo/audit', [
            'json' => [
                'target_url' => $url,
                'analysis_depth' => 'comprehensive',
                'include_recommendations' => true,
            ]
        ]);
        
        return json_decode($response->getBody(), true);
    } catch (\Exception $e) {
        error_log("SEO API error: " . $e->getMessage());
        return null;
    }
}

// Usage

echo "SEO Score: " . $audit['score'] . "\n";
echo "Recommendations: \n";
foreach ($audit['recommendations'] as $rec) {
    echo "- " . $rec['title'] . "\n";
}

function callApi($endpoint, $data = []) {
    $client = createApiClient();
    
    try {
        $response = $client->post($endpoint, [
            'json' => $data,
            'timeout' => 30,
        ]);
        
        return [
            'success' => true,
            'data' => json_decode($response->getBody(), true),
        ];
    } catch (\GuzzleHttp\Exception\ServerException $e) {
        return [
            'success' => false,
            'error' => 'Server error: ' . $e->getResponse()->getBody(),
        ];
    } catch (\GuzzleHttp\Exception\ClientException $e) {
        return [
            'success' => false,
            'error' => 'Client error: ' . $e->getResponse()->getBody(),
        ];
    } catch (\Exception $e) {
        return [
            'success' => false,
            'error' => 'Request failed: ' . $e->getMessage(),
        ];
    }
}

function rateLimitedRequest($endpoint, $data = []) {
    static $lastRequestTime = 0;
    $minInterval = 1; // seconds between requests
    
    $currentTime = microtime(true);
    $timeSinceLast = $currentTime - $lastRequestTime;
    
    if ($timeSinceLast < $minInterval) {
        usleep(($minInterval - $timeSinceLast) * 1000000);
    }
    
    $lastRequestTime = microtime(true);
    return callApi($endpoint, $data);
}

Additional best practices:

• Cache responses when possible to reduce API calls
• Use environment variables for sensitive credentials
• Implement exponential backoff for retries
• Monitor API usage quotas
• Handle timeouts gracefully