I should think that there should be a better way (e.g. port knocking, instructions for manually correcting the URL that cannot easily be automated, additionally supporting alternative protocols, etc).