In many circumstances, it is quite challenging for researchers to look for papers that match their interests when searched from databases. More often, papers which have been searched are perused individually to precisely identify the relevant contents. In this paper, we present a prototype to test our work-in-progress of an agent-based text mining algorithm that extracts and identifies the context of potentially relevant papers published in the WWW. The prototype involves an interface that enables a user to test the algorithm and view the results. The interface links to an abstract of a paper that the algorithm mines. We conduct tests on three abstracts of selected papers from the literature. We conduct the tests on various threshold settings that filter unnecessary data. The results show that the algorithm successfully identifies the contexts of the tested abstracts with thresholds of between 30%-60%.