An MVP isn't a small product. It's the shortest way to find out whether your users want it.
One journey, one user, one promise
A first version answers a single question: does this person, with this problem, use this product? Anything that doesn't help answer it can wait. So scoping starts on one page: who uses it, what costs them time, what needs to change, and how we'll know it works.
What you keep
The main journey, from start to finish, actually working. Accounts and permissions if they are needed. What it takes to measure usage. And a clickable prototype beforehand: we validate the key screens before writing a line of code, because that's when changing your mind costs almost nothing.
What you cut
Everything in the “not in V1” column: e-signatures, multiple currencies, settings nobody has asked for yet, rare cases you can handle by hand at first. They come back if users ask for them, and only then.
What you don't sacrifice
The data model: it decides whether the product will still stand in two years. Access security: everyone sees exactly what they should see, and the database enforces it. A real production launch: a domain, environments, a clean deployment. Cut the scope, yes. Skimp on the foundations, no.