Linktree scraper

tree-scraper.php

<?php

function scrapeLinktreeProfile($profile) {
    $url = 'https://linktr.ee/' . $profile;

    $ch = curl_init();
    curl_setopt($ch, CURLOPT_URL, $url);
    curl_setopt($ch, CURLOPT_RETURNTRANSFER, true);
    curl_setopt($ch, CURLOPT_USERAGENT, 'Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/58.0.3029.110 Safari/537.3');
    $html = curl_exec($ch);

    if ($html === false) {
        throw new Exception("Failed to fetch URL: $url");
    }

    $dom = new DOMDocument();
    libxml_use_internal_errors(true); // Disable libxml errors
    $dom->loadHTML($html);
    libxml_clear_errors();

    $xpath = new DOMXPath($dom);

    // Find the script tag containing the JSON data
    $scriptTags = $xpath->query('//script[@id="__NEXT_DATA__"]');
    if ($scriptTags->length == 0) {
        throw new Exception('Unable to find script tag with id "__NEXT_DATA__"');
    }

    $jsonData = $scriptTags[0]->nodeValue;
    $data = json_decode($jsonData, true);

    if ($data === null) {
        throw new Exception('Unable to parse JSON data');
    }

    // Extract information from the JSON data
    $formattedResults = [
        'title' => $data['props']['pageProps']['pageTitle'],
        'username' => $data['props']['pageProps']['account']['username'],
        'profilePictureUrl' => $data['props']['pageProps']['account']['profilePictureUrl'],
        'links' => array_map(function($link) {
            return [
                'title' => $link['title'],
                'url' => $link['url']
            ];
        }, $data['props']['pageProps']['links']),
        'socials' => array_map(function($social) {
            return [
                'type' => $social['type'],
                'url' => $social['url']
            ];
        }, $data['props']['pageProps']['socialLinks']),
        'raw' => $data
    ];

    return $formattedResults;
}

// Example usage:
try {
    $profile = 'linktr.ee';
    $result = scrapeLinktreeProfile($profile);
    print_r($result);
} catch (Exception $e) {
    echo 'Error: ' . $e->getMessage();
}
?>