That said, if you created a great tool perhaps it could eliminate the need to create a harness each time. However, I still feel this is most likely an instance of so-called "bike shedding".
You ask a good question at the end:
> do most people rely on other methods (Cypress, mocks,
and if you ask that of yourself I think you'll at least be able to narrow the scope of your idea. For example... why don't you just use Selenium to help recreate the test and save the 30-90 minutes? Perhaps this will help you find what the other tools are missing that will help differentiate yours. IDK
In any case, I don't want to discourage you. So, if you still love the idea I suggest you build a simple prototype for yourself. Then, see if you find it especially useful and pass it around for others to evaluate. It could turn out to be the greatest thing since sliced bread, you never know.