We implemented the Sift system to demonstrate the approach and to provide a testbed for the specific extraction methods proposed. The core of the system is a two-tier Python-based web application composed of three primary components. Two of these components, the browser extension and the topics dashboard, interact through web services with the backend server. Client-server communication between the tiers of the system is implemented via a REST interface over HTTP. In the case of the browser extension, these requests are made asynchronously via AJAX calls and return JSON objects, while the web interface uses traditional HTTP GET and POST requests for HTML documents.