Supabase is an excellent BaaS made in elixir if you want tips in that direction.

Is maintaining the Elixir backend your only problem? Is there any other issues related to Elixir that made you regret your decision? Because if there isn’t, then whatever backend you decide to have you will have to maintain it. Unless you can go full Firebase but that comes with its own limitations.

Also, if web is involved, Flutter Web is a bad, bad framework. For MVP you could maybe stick to familiar tech stack and do something like OnsenUI or Ionic, and then move to more native solution when you have some dedicated availability? I haven’t been on this space since 2017 but those were good enough back then.

Edit: I wouldn’t bet on LiveView for mobile apps just yet. You need to keep your options open and having an API is optimal for that. You can wrap a web view or move to native and even switch from flutter to compose etc if you have this option. With LiveView you might have to touch some backend code again to switch front end to native if (and with mobile, it’s more of a when than an if) needed.

How is Flutter/Supabase friendliness? I know it has some magic when Firebase is involved.

If you still have the Svelte code, live_svelte might be worth a look.

For me maintaining the server isn’t the issue, it’s having to manage both front-end and back-end codebases as a solo dev while also dealing with Flutter’s quirks. I’m also questioning if I can see myself coding long-term like this because it’s a lot of dev time to give people native apps like this, and it’s not even out. You can see that it looks meh, and the onboarding has a lot to get through. It’s difficult concepts to teach people on how to use the app unless they’ve gone to therapy, so I need to nail the onboarding.

Essentially there’s a lot of mental overhead for my small brain to upkeep with both client and server, and I’m not too content with the client

My hesitation is vendor lock-in when it comes to superbase, but it is really solid when it comes to DX. Had a shot at it for this app and it was smooth!

From what you are saying, the split brain between front and back is what is killing your productivity. It also looks like you want to go mobile first, so to me, going with Supabase or another BaaS is the way to go (option 1). LV is not mobile ready (yet) and I don’t think JavaScript is going to avoid the split brain (plus JS frameworks tend to be moving targets).

You should also question if it is worth going with Flutter or focusing on one of Android or iOS. Check the demographics you are targeting. Most likely both have a large user base that will be more than enough to validate your product.

It is best to have something out there that you can iterate later on than nothing. Good luck!

Very valid points. Thank you Jose for weighing in! You’re right

I can’t say much about flutter, but I can say about native android development, it is and was a big failure from the start, the abstraction level is too low, you need to understand a lot of android internals, the material is scattered and mostly consists of low quality answers on stackoverflow, the amount of scaffolding to get something going is insane, bad and opinionated practices that hug both old applications and new ones.

As much as I hate js, I would go almost always for a web application instead of a native one, solely from upfront investment you have to make.

Well, I will share a short summary of a journey I went through myself since your current conflicts are very related to the ones I had.

I have several app ideas (web & mobile), and started to try different stacks to see “what they taste like”.

I knew Elixir a bit, so I tried:

On mobile: Flutter, Swift & SwiftUI, Kotlin & Jetpack Compose.

On the backend: Go and Rust.

I work as Frontend/JS Developer, so I knew what was available in the frontend web space.

After several months of experiments, what I concluded was…

As a solo developer, creating an indie business, if I want to:

• create a mobile app, I would create a native iOS app with Swift & SwiftUI.
• create a web app, I would choose Elixir (with Phoenix and LiveView).

And the most important part of the message. Why would this be my choice?

The main reason: cohesive and convergent ecosystem. This is the most important thing a solo dev must have in order to have a chance to compete with a company with more resources IMO.

Both in Swift and Elixir ecosystems I felt that with very few tools I could create really solid applications, that in other ecosystems it would be nearly impossible to accomplish by myself with the same quality, velocity, and scope.

So my advice to you is, and what I am doing myself: do not try to win on all platforms at once as a solo dev. Choose one and nail it

If you go the mobile route, focus on the client and use a BaaS (in iOS you even have CloudKit “for free”).

If you go the web route, well, Elixir is my current bet.

I was really invested in mobile apps but in the end, I chose to go all-in in Elixir since it will give me more possibilities to create different types of apps in a high-level quality.

I also love functional programming, so it is a blessing to have this settled in the level of the language, without arguments .

ps: liveview native is very interesting since it is the only cross-platform framework AFAIK that uses the same building blocks that native frameworks use, so it is a better long-term investment knowledge-wise IMO. Definitely, I’ll try it in the future.