Simply enter (or paste) a valid URL for a publicly accessible web page. Once you click annotate, Wordbot will go fetch that URL in the background, and read it to determine what text is useful.
You can paste text into the text box, and Wordbot will read the text, analyze it, and list the important terms. Note that you can paste HTML, but the tags will simply be removed, unlike the URL retriever, which also attempts to remove whole chunks of the document which aren't part of the main content.
It's best to select a web page which focuses on a particular topic rather than a portal or gateway page. For example, if you want more information on the first article listed on the homepage of the New York Observer, you should paste in the link of the article itself, not the link to the homepage.
You can browse the various sections of the document results by clicking the tabs on the the top of the page. You can also choose to enable any term that Wordbot has selected from the document by clicking it in the tag cloud on the right side of the page. Red terms are considered in searches, clicking them will disable them. Black terms are not considered in searches, clicking will enable them. Wordbot automatically selects the highest scoring term from the document.