Migration to AWS from Lovable
Discussion
Not sure whether to put this here or in Business.
In my spare time I’ve developed a multi-tenant SaaS application in Lovable. Repo is fully synced with GitHub. Front end is hosted on Lovable. UI is Lovable/Tailwind. Code is React. Back end is Supabase (Pro account). Quite a few Edge Functions in Supabase. Uses Resend for mail, Puck for other stuff. It has multiple domains and sub-domains based on what I need it to do.
It’s more than an MVP now - it’s fully working and integrated with Stripe and GoCardless sandbox accounts. At this point I have no customers, just a bunch of test users giving me feedback. The UI could with refinement but that’s for another day.
The question is this - do I take the plunge now and migrate it to AWS or do I keep it on Lovable and sign up some real customers? (maybe one or two friendly ones either free or heavily discounted). If I do the latter (Lovable), how hard is it to migrate from Lovable to AWS when the application is live? The Supabase data would remain intact but I am concerned that the auth functionality might be tricky to migrate with live users. I keep flipping between testing it with a live customer in Lovable or making it robust and scalable now in AWS. One of my concerns is that there might be some lovable calls in the code that I haven’t spotted and which will mess up when I migrate. Yep, pitfalls of AI coding.
If this all sounds like nonsense and jargon, please ignore. But if you can help a non-dev navigate through this I’d be grateful for any pointers. I’m coming at this with a mechanical engineering mindset, not a tech one. But I think I understand enough to listen to advice! Please be gentle with me if I’ve got any terminology wrong. I’ve been learning how to put this stuff together since May and I’ve got my head around most of it (including RLS policies in Supabase) but I probably come across as a bit of an amateur if this is your day job.
Cheers.
In my spare time I’ve developed a multi-tenant SaaS application in Lovable. Repo is fully synced with GitHub. Front end is hosted on Lovable. UI is Lovable/Tailwind. Code is React. Back end is Supabase (Pro account). Quite a few Edge Functions in Supabase. Uses Resend for mail, Puck for other stuff. It has multiple domains and sub-domains based on what I need it to do.
It’s more than an MVP now - it’s fully working and integrated with Stripe and GoCardless sandbox accounts. At this point I have no customers, just a bunch of test users giving me feedback. The UI could with refinement but that’s for another day.
The question is this - do I take the plunge now and migrate it to AWS or do I keep it on Lovable and sign up some real customers? (maybe one or two friendly ones either free or heavily discounted). If I do the latter (Lovable), how hard is it to migrate from Lovable to AWS when the application is live? The Supabase data would remain intact but I am concerned that the auth functionality might be tricky to migrate with live users. I keep flipping between testing it with a live customer in Lovable or making it robust and scalable now in AWS. One of my concerns is that there might be some lovable calls in the code that I haven’t spotted and which will mess up when I migrate. Yep, pitfalls of AI coding.
If this all sounds like nonsense and jargon, please ignore. But if you can help a non-dev navigate through this I’d be grateful for any pointers. I’m coming at this with a mechanical engineering mindset, not a tech one. But I think I understand enough to listen to advice! Please be gentle with me if I’ve got any terminology wrong. I’ve been learning how to put this stuff together since May and I’ve got my head around most of it (including RLS policies in Supabase) but I probably come across as a bit of an amateur if this is your day job.
Cheers.
Good point. I’m happy that it all works as-is but I don’t know if it’s a viable business proposition. If it proves to be viable (i.e., customers like it) I’ll move to AWS. If they don’t, it’s just more effort late at night when I get time to do this stuff. The question is - should I put the effort into AWS now (which is on top of the 250 hours I’ve spent on it at night and weekends) or give it a whirl using lovable just now? If migrating from lovable to AWS is a royal pain in the butt then I’ll do it now. Hope that clears things up.
(I have no knowledge on Lovable)
If your doing this to build a successful product / business, you should be thinking about attracting users first and foremost.
If you want to build a product your proud of / demonstrate engineering mastery even if no-one uses it (and there's nothing wrong with that!) then fix it first.
w1bbles said:
The Supabase data would remain intact but I am concerned that the auth functionality might be tricky to migrate with live users.
It's admirable to avoid frustrating your users, but that can be a problem for when you actually have some!If your doing this to build a successful product / business, you should be thinking about attracting users first and foremost.
If you want to build a product your proud of / demonstrate engineering mastery even if no-one uses it (and there's nothing wrong with that!) then fix it first.
Jakg said:
It's admirable to avoid frustrating your users, but that can be a problem for when you actually have some!
If your doing this to build a successful product / business, you should be thinking about attracting users first and foremost.
If you want to build a product your proud of / demonstrate engineering mastery even if no-one uses it (and there's nothing wrong with that!) then fix it first.
I think you are outlining both sides of my argument in one post! If I want to attract users I’ll run with the lovable implementation and see if the product is good. I’m on this. But you also say ‘fix it first’. So… MVP on lovable or next level dev on AWS? I’m all ears.If your doing this to build a successful product / business, you should be thinking about attracting users first and foremost.
If you want to build a product your proud of / demonstrate engineering mastery even if no-one uses it (and there's nothing wrong with that!) then fix it first.
From reading your OP, I get the impression the most of the development has been done using AI?
In which case I would be inclined to leave it as it is for now. It works to a point and you're familiar with Lovable.
There will be bugs, issues, and vulnerabilities that you haven't yet discovered. AI generated code will be full of them.
You're still in development, so keep your development environment as familiar as you can and avoid adding any more complexity than absolutely necessary. Don't add the headache of dealing with AWS to the headache of fixing codebase bugs for paying customers.
It is fashionable to worry about scalability before a project has a single customer. Don't. Concentrate on getting this thing paying for itself first. Then concentrate on developing features actual paying customers need. Then worry about scale when you have enough customers to warrant it.
In which case I would be inclined to leave it as it is for now. It works to a point and you're familiar with Lovable.
There will be bugs, issues, and vulnerabilities that you haven't yet discovered. AI generated code will be full of them.
You're still in development, so keep your development environment as familiar as you can and avoid adding any more complexity than absolutely necessary. Don't add the headache of dealing with AWS to the headache of fixing codebase bugs for paying customers.
It is fashionable to worry about scalability before a project has a single customer. Don't. Concentrate on getting this thing paying for itself first. Then concentrate on developing features actual paying customers need. Then worry about scale when you have enough customers to warrant it.
jagnet said:
From reading your OP, I get the impression the most of the development has been done using AI?
In which case I would be inclined to leave it as it is for now. It works to a point and you're familiar with Lovable.
There will be bugs, issues, and vulnerabilities that you haven't yet discovered. AI generated code will be full of them.
You're still in development, so keep your development environment as familiar as you can and avoid adding any more complexity than absolutely necessary. Don't add the headache of dealing with AWS to the headache of fixing codebase bugs for paying customers.
It is fashionable to worry about scalability before a project has a single customer. Don't. Concentrate on getting this thing paying for itself first. Then concentrate on developing features actual paying customers need. Then worry about scale when you have enough customers to warrant it.
Thanks - that's a good perspective.In which case I would be inclined to leave it as it is for now. It works to a point and you're familiar with Lovable.
There will be bugs, issues, and vulnerabilities that you haven't yet discovered. AI generated code will be full of them.
You're still in development, so keep your development environment as familiar as you can and avoid adding any more complexity than absolutely necessary. Don't add the headache of dealing with AWS to the headache of fixing codebase bugs for paying customers.
It is fashionable to worry about scalability before a project has a single customer. Don't. Concentrate on getting this thing paying for itself first. Then concentrate on developing features actual paying customers need. Then worry about scale when you have enough customers to warrant it.
If you moved away from lovable would you be capable of making any changes required?
It does seem a good way to lock people into an eco system, by making them reliant on the ai. Otherwise its just code and it should run anywhere.
I would have a look at what it requires to get it running in AWS. They are not going anywhere. I had to Google lovable to find out wtf they are
It does seem a good way to lock people into an eco system, by making them reliant on the ai. Otherwise its just code and it should run anywhere.
I would have a look at what it requires to get it running in AWS. They are not going anywhere. I had to Google lovable to find out wtf they are
Gassing Station | Computers, Gadgets & Stuff | Top of Page | What's New | My Stuff