How to Crawl Web Pages with PHP



Uwe Hunfeld provides an object oriented library called PHPCrawl available at This class can be used to crawl web pages with many different parameters. It also allows you to process each page and do what manipulation or scraping you need to do. You accomplish this by overriding the base class and implementing your own functionality in the handleDocumentInfo() and handleHeaderInfo() functions. Use the code below as an example of how to create your own web crawler. If you want to learn how to parse the HTML DOM and extract things like links and headings, check out the post on How to Parse HTML DOM with PHP.

set_time_limit(1000); // Set execution length in seconds

// Extend the provided base class and override the handler functions
class NanoCrawler extends PHPCrawler
    // Process the document contents in $DocInfo->source here
    function handleDocumentInfo($DocInfo)
        echo "Page requested: ".$DocInfo->url." (".$DocInfo->http_status_code.")\n";
        echo "Referer-page: ".$DocInfo->referer_url."\n";
        if ($DocInfo->received == true) {
            echo "Content received: ".$DocInfo->bytes_received." bytes\n";
        } else {
            echo "Content not received\n";
        echo "Links found: ".count($DocInfo->links_found_url_descriptors)."\n\n";


    // Process the headers like http_status_code, content_type, content_length,
    // content_encoding, transfer_encoding, cookies, source_url
    function handleHeaderInfo($header) {

// Instantiate the new custom crawler class we defined
$crawler = new NanoCrawler();

// Set rules and params

// Ignore links to pictures, dont even request pictures
$crawler->addURLFilterRule("#\.(jpg|jpeg|gif|png)$# i");

// Store and send cookie-data like a browser does

// Set the traffic-limit to 10 MB  (in bytes)
$crawler->setTrafficLimit(1000 * 1024 * 10);

// Set user agent. It's not polite to lie about your user agent.
// If you are creating a bot or crawler it is good to set the user agent as
// something unique that includes a way to contact like website.
// This allows people to report any problems or block your user agent
// if it causing problems. There are other situations where a website
// will reject anything but a familiar user agent. Sometimes just putting
// "firefox" as a user agent will bypass filters.
$crawler->setUserAgentString("Mozilla/5.0 (Windows NT 6.1; WOW64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/34.0.1847.116 Safari/537.36");

// Delay between requests in seconds

// Max number of pages to crawl
$crawler->setPageLimit(5, true);

// Skip items listed in robots.txt (or not)

// Follow redirects

// Run it. May take a while.

// Output crawl report
$report = $crawler->getProcessReport();
echo "Summary:\n";
echo "Links followed: " . $report->links_followed . "\n";
echo "Documents received: " . $report->files_received . "\n";
echo "Bytes received: " . $report->bytes_received . " bytes\n";
echo "Process runtime: " . $report->process_runtime . " sec\n";
echo "Memory peak usage: " . (($report->memory_peak_usage / 1024) / 1024) . " MB\n";