This is an incomplete project. Feel free to fork and develop.

require_once('Unhomoglyph.php');

Unhomoglyph::init('homoglyphCharmaps/extended.php');

$str1 = 'google.com';
$str2 = 'gοοgle.com'; // o's are actually u+03BF (lowercase omicron)
if( $str1 !== $str && Unhomoglyph::skeleton($str1) === Unhomoglyph::skeleton($str2) ) {

	echo "WARNING! $str1 looks like $str2 but they are NOT the same";

}

Full lookalike/homoglyph character map is in homoglyphCharmaps/extended.php. It's organized into unicode blocks which may help you if you only need to focus on certain ranges. A TODO is to automatically break this into separate files using code generation and only include blocks as needed.

The base of the character map is originally based on https://github.com/nodeca/unhomoglyph which is based on http://www.unicode.org/Public/security/latest/confusables.txt. However, this map left a lot to be desired, and I've gone through the first 5,000 unicode characters or so manually and updated the mapping. Unicode has over 150,000 characters at the time of this writing. See a Unicode Block that needs improvement? Donate it with a pull request!

Unhomoglyph::renderCharmapTable() - Renders a table of character => mapped character Unhomoglyph::exportUpdatedOrganizedCharmap() - Sorts and re-organizes extended charmap, generating updated return array
Unhomoglyph::exportInverseGroupedCharmap() - Generates an array of skeleton character => homoglyphs
Unhomoglyph::wikipediaUnicodeBlockTableParserApp() - Tool to generate Unhomoglyph::$blockRanges from Wikipedia