Roistr is a project that’s smaller than our typical API feature. It shows the falling barrier to entry in the API world even in a the demanding academic area of natural language processing and semantic analysis. The primary Roistr API function is a “best match” engine. It offers an interesting use case to demonstrate the relevance of the service: take a Twitter username and use tweets from this account to find content most relevant to the user’s interests.
Using a Twitter account to sign in at a new site has become a common experience on today’s web. I thought of this as mainly a convenience to the user. What I didn’t realize is that in addition to the benefit of not needing to create yet another username and password, the website can use public profile information on twitter to customize the users’s experience of the site. I typically thought of this as more of a feature of Facebook sign-in, when you just assume that you’re giving away all you’re personal details to some anonymous party.
The Roistr API is RESTful and is offered with a client implemented in Python. This site seems very new since there are lots of plans for expansion. It seems like each passing week brings another player to the natural language and semantic analysis space. Currently there are more than 80 semantic APIs in our index. Will Roistr gain staying power? Stay tuned to find out. The service kind of reminds me of Text-processing.com, which we profiled back in April 2011.