The Wiki contains about a 1000 pages, and the crawler examines approximately 1 page / second so it should take around 20 minutes for the crawl to finish. For the rest of the screencast, I will use another application, where I have already finished crawling the Seinfeld wiki.
Now, the easiest way to get started with the SDK is to download the minified version of mindmeld-2.0.js from here. You can also check-out the public mindmeld-js-sdk repository here. Lastly, you can install the SDK as a bower component like this: bower install mindmeld-js-sdk.
The first thing we need to do to start using the SDK is call MM.init with a config object specifying your app id and an onInit callback. Get your app id from the developer console and add it as a parameter in your config. In our onMMInit function, let's log a message and display the SDK version.
To start making calls to our API, we need a user token. The API supports two mechanisms for obtaining a token with user privileges: simple user authentication. We'll be obtaining a token via simple user authentication here, which requires the app secret. Simple user authentication allows you to obtain a user token by submitting valid administrator credentials along with a set of credentials that uniquely identify the users of your application.
Let's create our credentials object so we can get a new user token. Obtain your appsecret from the developer console page. All entity extraction and document ranking happens inside of a MindMeld session. Now, let's create a new, private session, named 'seinfeld session.' Using this SDK, you create you own session to which you can later add text entries and fetch relevant documents.