What tech stack should a solo founder use?

Use the dullest stack you already know, write it on one card, and build everything on it. A new product should never also be a new stack, because one person can't afford to learn a framework and find customers in the same month.

The card

My card has 6 lines: the language, the framework, the database, the host, the payment provider and the email service. I wrote the first version in my first year, the week my first product died. I've changed it twice in 9 years, both times in January.

Since my third year I've only allowed changes in January. Ideas for a new tool go on a list. In January I look at the list, and most years I leave the card alone. (The card usually wins.)

Why the first product taught me this

My first product was a booking tool for music teachers. I built it on a framework I'd wanted to try for a year and a database that was fashionable at the time, on the strength of articles saying the ordinary one would fall over under load. It took 6 weeks. It had no customers to put any load on anything.

The stack was good work. Nobody ever saw it. It was the one part of the product that didn't matter, and it took most of the time.

What boring buys you

Speed, mostly. When you know the stack, a small product can be live in a week, because the only new thing is what it does.

It also buys you sleep. A stack you know fails in ways you've seen before. And when you run more than one product, a shared stack means one way of deploying, one backup job and one set of surprises to know about, instead of one per product.

When to break the rule

When the product genuinely can't be built on the card. That's rare. "I'd like to learn it" isn't a reason; it's a hobby, and hobbies make excellent hobbies.

Ship Alone sets out the dull stack in full. The Second Product shows what one stack does for 5 products.