The "usefulness" of a critical apparatus depends both on the editor's judgment of what to include or exclude and also on a given reader's needs, which may or may not align with the editor's critical principles. The challenge, then, is to make the critical apparatus flexible – to allow the reader to change the level of detail presented in the apparatus, on demand. In this paper, I will present a hypertext edition and an open-source software platform currently in development that performs automatic collation on demand and generates an apparatus for a base text and a set of variant texts, with constraints on the level of detail to be included in the apparatus that can be adjusted by the reader. These constraints are essentially the reader's text-critical principles, expressed as parameters that guide the collation algorithm. I will consider the specific challenges and solutions of such a platform as they relate to the presentation of Sanskrit texts as well as the general scholarly aims that I hope to achieve by producing such a work.