I am currently writing a thesis that is meant to revolve around an experimental software implementation I am also working on. I usually think of theses as being scientific in some way, but there is limited experimentation I can do with my project. I know I can contribute are my observations on how this software was built and what its prowess are and what it limitations are, but I want to avoid it sounding like an infomercial. How should I approach writing a thesis like this?
Start by talking to your supervisor about it. They ultimately need to approve it, so you need to know their criteria for what they'll accept.
But generally speaking, there are several possibilities:
If your software is intended to show that a certain theoretical technique works in practice, then you should also create some test cases on which you can test whether your software in fact does that.
If your software is intended to be an improvement compared to an existing product, then you need to find some benchmark which you can compare them both on.
If your software is intended to solve a particular problem or work on a particular data set, do experiments to show that it indeed works on it.
Ask yourself: how would you test if this software works correctly?
What is your software for? What problem does it solve? Why is it reasonable to solve that problem in the way you did?