I agree here, billing is not something I would put in an MVP. Odds are no one will sign up, or so few people will sign up that that billing code will just go to waste.

Bill manually for the first few customers.

I will second this. I recently replaced Google maps with Leaflet.js. And as soon as I had a beta ready, and I had updated sites with the new code, I found a better way to render the maps using Mapbox's own js libraries.

But it doesn't matter to anyone using it. They are just glad to be rid of Google maps, right now. If there's another small hurdle, they don't mind. And if I make improvements down the road, all the better.

You can always add the "convenience" of in-app billing later. A nice upgrade that will make your users happier. :)

Or, as I have often found, you may decide to _never_ put billing into your app. Why? Because you may learn that your users' are just fine with paying via your website instead. But you would only find this out by launching now, and talking with your clients.

For the first couple of customers just use Square / Stripe and do manual billing and invoicing.

I wanted to answer with the same theme - don't wait to have billing perfected on all edge cases, focus on covering the normal cases and release. You can always go back and fix those edge cases later, if they even ever happen to you.

More in general, when you have a mountain that looks too high to conquer, set a partial goal that looks achievable and only focus on that one, without thinking about the end goal until you reach the partial one. This helps a lot in my experience.