How to identify the client is a search robot?

I have built my entire site using AJAX (indeed it's GWT). I have also implemented AJAX crawling proposed by Google. However, after the implementation, I found that neither Yahoo , Bing, nor Baidu implemented that scheme!

I'm wondering if there is a way to identify the web client is a search robot. If they are, they will be shown the HTML snapshot I created.

It will be best if I can identify them in APACHE level, then I can just do a mod_rewrite. But it's still ok if I can do that in PHP or GWT.

It's quite tricky as there are so many different search engines. I guess this can't be done in a one liner. If PHP is ok, I suggest using php_browscap.ini (available for different languages). This can be used with PHPs get_browser()-function, for example.

Then you can check against all different browser stuff and of course Bots.

In my own code snippets I use the following function (based on the mentioned browscap) to get the appropriate infos easily in my code:

function _browser($a_browser = false, $a_version = false, $name = false)
{
 $browser_list = 'msie firefox konqueror safari netscape navigator opera mosaic lynx amaya omniweb chrome avant camino flock seamonkey aol mozilla gecko';
 $user_browser = strtolower($_SERVER['HTTP_USER_AGENT']);
 $this_version = $this_browser = '';

 $browser_limit = strlen($user_browser);
 foreach ($this->_w($browser_list) as $row)
 {
     $row = ($a_browser !== false) ? $a_browser : $row;
     $n = stristr($user_browser, $row);
     if (!$n || !empty($this_browser)) continue;

     $this_browser = $row;
     $j = strpos($user_browser, $row) + strlen($row) + 1;
     for (; $j <= $browser_limit; $j++)
     {
         $s = trim(substr($user_browser, $j, 1));
         $this_version .= $s;

         if ($s === '') break;
     }
 }

 if ($a_browser !== false)
 {
     $ret = false;
     if (strtolower($a_browser) == $this_browser)
     {
         $ret = true;

         if ($a_version !== false && !empty($this_version))
         {
             $a_sign = explode(' ', $a_version);
             if (version_compare($this_version, $a_sign[1], $a_sign[0]) === false)
             {
                 $ret = false;
             }
         }
     }

     return $ret;
 }

 //
 $this_platform = '';
 if (strpos($user_browser, 'linux'))
 {
     $this_platform = 'linux';
 }
 elseif (strpos($user_browser, 'macintosh') || strpos($user_browser, 'mac platform x'))
 {
     $this_platform = 'mac';
 }
 else if (strpos($user_browser, 'windows') || strpos($user_browser, 'win32'))
 {
     $this_platform = 'windows';
 }

 if ($name !== false)
 {
     return $this_browser . ' ' . $this_version;
 }

 return array(
     "browser"      => $this_browser,
     "version"      => $this_version,
     "platform"     => $this_platform,
     "useragent"    => $user_browser
 );
 }

function _w($a = '')
{
    if (empty($a)) return array();

    return explode(' ', $a);
}

Examples:

/*
// Examples

echo '<pre>';
print_r(_browser()); // return array of browser data

var_dump(_browser('firefox')); // return true if using firefox
var_dump(_browser('msie', '>= 7.0')); // return true if using IE 7.0 or above else false
var_dump(_browser('firefox', '< 3.0.5')); // return true if using below firefox 3.0.5 (can check minor version)
var_dump(_browser(false, false, true)); // return string of name of browser and version

// To check if Gecko browser is used
var_dump(_browser('gecko'));

// version_compared function is used so you can use the same operator syntax
var_dump(_browser('firefox', 'le 1.5'));

echo '</pre>';

To do it with Apache you can use a rewriteRule together with a RewriteCond on %{HTTP_USER_AGENT} .

The RewriteCond accepts RegExp, so you have to use a pattern there, which matches all those bots(you should find informations how to build the pattern on the page linked by acme).

But be careful: Search-engines may penalize pages that deliver different contents to the bots and to the others clients.

If you use PHP, you can check $_SERVER['HTTP_USER_AGENT'] against the string of search engine User Agent string -- some details are found here: http://en.wikipedia.org/wiki/Web_crawler

You can also use a Firefox plugin to see how the search engines will see your page: http://chrispederick.com/work/user-agent-switcher/