The main heart of the project is a client side app written in Scheme/WASM and some JS Web Components and plumbing. In the main use case (people training on it, it's an ed tech thing), users will load the client side app, load saved exercises, and save their custom exercises. So most of the time, they aren't doing anything to the server. But every few minutes, they might load a new exercise or save, requiring either a server side refresh (the exercises are done in HTML scaffolds); a json load; or a json save. (To begin with, these will be presisted as JSON blobs in PostGres, which could change if I ever start to outgrow a simple database.)

Sign up, account handling, payments, etc is handled in a Django app because it's bog standard and generates such low traffic that I will prioritize buy over build for that part. In day-to-day use, users will never go there so it really doesn't need to be scalable and I have urge to write my own Stripe handler and so on.

The catch is, I want to keep this dead cheap per user (not worried about initial cost). The app has been designed to ensure the DB grows very little per user (there is a cap to what they can save). So I want this part of the app to be able to serve as many users as possible on cheap, reliable machine(s) that I can sign up for and basically forget about.

Before people say "just use Django".... we will likely launch on Python only, but I definitely do not want to be only running Python if we achieve some magical viral marketing success as having to scale Python up rapidly is not a good place to be. Getting this small part ready on other tech is likely to take less than a week and will be worth it for the peace of mind alone. (I work in tech diligence, have done Python professionally for a long time, and have seen the scars.)

Hoping people with real world scaling experience can speak to this case. I could happily do this in any of Clojure, Elixir, or Common Lisp and wouldn't be bothered about the language.

thanks!