Techniques used include:
SprySpell is, as far as I am aware, completely unique in its application. Basically it is a spell-checker for web-sites. But rather than spell check the raw code, which most people can't read (and understandably so) SprySpell renders the site in question in the web-browser as normal, but it highlights words that it doesn't know and offers corrections.
SprySpell works in a multi-stage process: firstly SprySpell pulls the web-page off the target server using curl, and is then processed by a PHP script. The script will follow any 300 series server messages to find the correct document (ie www.sprymedia.co.uk actually goes to www.sprymedia.co.uk/index.php), and when it finally gets the document it parses through the html. Code, scripts, flash and so on are all stripped out, leaving just the words to be spell checked. The spell checking it done by interfacing with the excellent PSpell library on the SpryMedia server and SprySpell injects code into the document to offer interactivity and suggestions for word corrections.
The basic service offered on the SprySpell web-site [www.spryspell.com] will simply re-render the web-page with any corrections made, however, it is possible to install SprySpell on a web-site and have it interface with SpryPanel for spelling corrections there, and also with html files (requires ftp write access) - or a mix of the two.
SprySpell generally works very well, however there is the odd situation where it trips over code, particularly either Javascript in the body section, or bad coding practice. As such SprySpell is software which I make incremental updates to whenever I come across a site which doesn't work.