What's your favorite backend for building a MVP?
Let's say you are building a minimum viable product (MVP) and you have chosen React or Flutter as your cross-platform framework for the frontend. You don't want technical debt. So you want to stick with the technology stack.
Trending on Indie Hackers
Django
For a quick and dirty MVP:
Next.js + Prisma + Supabase + NextAuth
If SEO is not important and I've got more time:
React + Django DRF + Redis +Celery
same here: django, ninja, redis, celery + bootstrap
Any resources (or maybe a brief explanation) on why the second stack isn't good for SEO? I'm clueless when it comes to SEO 😅
I would use the second stack to build SPAs (Single Page Applications). I encourage you to google "SPA vs SSR for SEO" (server side rendering), you'd get a much more thorough explanation. But in short: search engine crawlers are smart enough to index SPAs pages (where HTML is rendered on the client after the JS is loaded) but a couple factor still negatively impact the indexing. Server side rendering is always the preferred solution when SEO matters (like indiehackers.com which pre-renders all of its pages on the server).
Why not go with Next.js for the frontend if you're already using React, that way you can selectively choose which pages to prerender?
Can you help me understand the reason for using NextAuth and Supabase at the same timr? I'm not very familiar with Supabase, but I thought the whole point of it is that it makes auth easy and comes with a bunch of auth solutions? And NextAuth is an auth solution that comes with Next.js, right? So I think I'm misunderstanding something, why use two auth solutions, why not just use Supabase for auth, or NextAuth and Postgres/Prisma (without Supabase)?
I only use Supabase for persisting the app data into a PostgreSQL db and storing media files. I don't want my auth system to rely solely on Supabase.
So what I do is I connect NextAuth to Supabase via a Prisma Adapter. The Prisma Adapter then creates User, Account, Session and Verification Token tables (for a magic link auth system) in Supabase. The reason I do that is because it gives me more flexibility and options in case I'd like to migrate my data to another vendor/provider. If that happens, I would not have to recode the whole authentication system in my app.
Thank you very much for your reply, this is very helpful!
Could you also tell me - if you're not using Supabase for auth, then what is the benefit of using Supabase over just using Prisma and a PostgreSQL hosting provider (like AWS RDS)?
Also, do you find NextAuth nice to use? I remember trying it, and finding it very difficult to customize - I wanted my login/signup form to appear in a modal, instead of a separate page provided by NextAuth, and that was surprisingly difficult to do.
Also, do you have any open source code that shows an example of setting this stuff up?
Sorry for so many questions, I have a big project coming up that will use this stack, I need to make a bunch of decisions about setting it up, and any advice from a more experienced person would be extremely helpful!
As is said, this is my "quick and dirty" way of building an MVP. I don't necessarily need the most robust/future-proof tech for prototyping. Supabase is free and incredibly easy to get started with thanks to its GUI . This gives me some runway and facilitates operations until I can get traction and validate my idea.
In my case, I don't need to heavily customize the auth flow as I'm using a magic link to sign up and authenticate my users. Again, this would be for an MVP so all I want is keep it stupid simple (KISS)! Passwordless authentication is getting traction out there, including amongst big names in the tech industry. Perhaps this could be a solution worth exploring?
Frontend Masters recently published a 10 hours long workshop where they show how to use Next.js with Prisma, NextAuth and Postgres+Heroku. It's taught by a Netflix Engineer: https://frontendmasters.com/courses/fullstack-app-next/
I hope this helps!
Another good tutorial out there: https://alterclass.io/courses/build-a-fullstack-app-with-nextjs-supabase-and-prisma-322389284337222224
Rails
I use the old school (but solid) LAMP stack
Linux + Apache + MySQL + PHP
LAMP on Dreamhost. I am familiar with it.
No libraries or frameworks.
Can get a page up in a few seconds.
I use DigitalOcean to host. :)
Did you face any challenges in scaling?
Not really! it's a misconception.
PHP-MySQL is a powerful duo. A web app's performance and scalability totally depend on the database. MySQL is a pretty powerful DB to do most of the things that you want to do.
Flask 😍
This comment was deleted 2 years ago.
Django for me.
+1
NextJS + Supabase is my favorite combination. They handle almost all heavy lifting for an MVP for you (HTTPS, user registration, emailing users, analytics, app deployment, etc) and they both a have a generous free tier.
what do you use for authentication? Nextauth or supabase auth?
Supabase
Django, because i am good at it
Django
Great tool
NestJS for full-stack TypeScript apps. Lets you develop much of the app in the front-end and copy-paste it to a backend as the UI starts to stabilize
Angular + NestJS = ♥
I typically use Rails, and if I need a SPA front-end, I use InertiaJS.
https://inertiajs.com/
It works with Rails/Laravel along with React/Vue/Svelte.
I am working in a project with Rails+Inertia+Vite+Vue, the setup had a few hiccups, but now it is pretty pleasing. No APIs, no serializing/deserializing, no reasoning about state in the front-end, no GraphQl!!!! (thank the lord). The Inertia API takes a little bit to be learned but it works well. The usual disclaimer is that I have just a few weeks with the project and I am the only person working on it so far so YMMV.
I have been able to avoid SPAs altogether lately with Hotwire 😅
I find Stimulus too verbose, but I see your point. I dont mind (and even like) the front-end frameworks per se, Vue/React/Svelte, but I hate hate hate the rest of the ecosystem...routers, validators,stores, messages, mutations....blah blah BLEH
I use NestJS + PostgreSQL for all my backend work since its easy to transition to React with Typescript on the frontend
In my case, Firebase provides everything for running the business. The cost is still less than $1 until now. I have run it for about 3 years :|
Laravel for sure.
With the power of Vue and Inertia it becomes super easy to quickly create SPA's. And Laravel is still powerful when it comes to just RESTful microservice, even though there are better tools out there.
What would you say are better tools than Laravel for the RESTful microservices?
I would say there are two options for me:
I know that FastAPI (python) is unreal cool & easy thing, but haven't tested it yet.
Next.js not only frontend, with API Routes. Too too easy and fast.
I LOVE Next.js. Makes it super easy to make a full stack app
FastAPI + Jinja
Thanks for you comment. I want to know what technologies will survive in the next 5 years. Your input is valuable.
LAMP. Look up the Lindy effect.
Maybe if you don't need users...
Do you mean fastapi or jinja is bad?
Django has user support out of the box, while FastAPI needs some extra work
You should check out redwoodjs
Supabase is amazing too!
Laravel forever ❤️
Flask python
I think Node and Express are the best balance between ease of use/getting started and performance to improve upon when building the product from the MVP seed
I'm pretty deep into the Laravel ecosystem. Laravel, Vue/React, InertiaJs and DigitalOcean/Forge for dev ops.
My favorite backend for building a MVP is Python. I like Python because it's very easy to learn and use, and also has great documentation.
In my experience, Python has been the most reliable backend for building MVPs. It's easy to get started with, but also has a lot of flexibility when it comes to customization.
I would recommend that you try out some different backends before deciding which one is best for your needs.
Kirby CMS or plain HTML, CSS + JS
Spring + React Admin + keycloak.
Mainly because that is what I know and that is what i code when freelancing to pay the bills and I feel at home with Java.
Is Java actually good to quickly ship new products from your point of view?
Difficult to say, it certainly requires a learning curve. It consumes more resources than golang etc so it costs much more to host. It is not that fast as node.js. But, that is what I know, it has every lib you can think off, great support from the community and great documentation. It is great for very large monoliths and microservices since java is very organized.
Specifically, security wise. It is easy to get a report on vulnerabilities and upgrade but that is not that important for an mvp.
I think that sticking to what you know is ok for an MVP.
"Not as fast as node.js" What do you mean by that?
Performance wise. I have seen some articles lately. Seems like nodejs is really optimized from the get go. I think to beat it youll need something like java vertx but believe me it is less fun than spring. However, usually unless speed really matters, spring java is ok for FE responses. Not the best, probably wont solve your 10k connections etc. But lets face it, just add 2 more machines and youll get back to ok. and if it is easier to maintain than who cares.
Well, NodeJS stands out if the app needs a large number of real-time connections. When we need a non-blocking technology on the backend. Though Java has Spring Reactor framework for that. A generic statement that NodeJS is faster than Java is incorrect on many levels. Java is as quick as NodeJS. Every technology has specific use cases.
Not java, spring as it comes by default with tomcat vs the default nodejs. Obviously, any language optimized enough will have pretty much the same amortized performance. The difference is how hard will it be to get to the same performance when you start to tweak the defaults.
If you were to hire a developer which technology would you pick? To paraphrase, which framework/stack you wished you knew.
Maybe something like next.js and typescript.
Or beego.
I once did something with python and flask and simple html really quick.
But, if this is not an mvp but production i think java with react is a great choice for a closed platform. The only trouble is seo on the react side and i think this can be solved by rendering static pages and serving with nginx or something like nodejs. The main problem is getting the same experience on the closed and open platforms. However, if SEO is no trouble then java is good just costs more to host.
AWS CDK
I'm a big fan of AWS CFN Templates and SAM, but CDK looks really cool.
Have you used CFN and/or SAM? How do they compare to CDK?
CDK actually generates a CFN template when you run cdk deploy and pushes that to AWS. No more dealing with YAML and you get all of the nice, obvious things you get from a general programming language. You can also push your CDK stack to SAM for local development.
I use CDK every day now.
You go straight to cdk over using amplify?
Yes absolutely. I've used both and at this point Amplify is deprecated in my mind. CDK can do everything Amplify does, plus more. You can even handle Amplify resources from CDK.
Not needing to monkey around in the console as much is a big plus too.
I am absolutely a cheerleader for CDK. Check out the docs and prepare for a new, beautiful world where all your resources are defined in general programming language.
I've used Amplify for hosting front end react apps and really like it, but never for backend/full stack stuff.
Have you used it, what's your opinion?
You guys should checkout appwrite.io
Recently used for an MVP, so far so good
Symfony (Laravel-ish)
Are you using "pure" Symfony or maybe ApiPlatform?
For the main project right now, "pure".
Next.js or Remix
Thanks for you comment. I wish I had added more Node JS based technologies in the poll. But, there are so many of them .. ;-)
Whatever you feel comfortable with. MVP is about speed. Use what you can get up and running in the smallest amount of time.
Actually, I am doing market survey. I want to know what would be most useful to as many indie hackers out there.
You could load up a couple hundred indie projects you find on this site and check for "Powered-By" HTML headers. A lot of frameworks send them.
NestJS
Thank you. I appreciate your comment. This will help me in deciding the tech stack.
My go-to is Laravel, but I always felt it was over-engineering for an MVP. Recently started working with Xano and that platform helps you setup right out of the box. Plus after you figure your way around it's No Code interface, the whole process is a breeze.
NestJS + Angular or Supabase + Angular :)
RedwoodJS
Sitefox is a ClojureScript back end framework I built that is optimized for indie hackers who need to ship fast. That's what I use of course!
For a React Frontend I'd use Thin Backend (https://thin.dev/)
No love for Rails?
exactly
Flask, Postgres, Redis, Celery and Firebase for Auth.
How do you make Firebase Auth work with rest of the stack?
I suppose Firebase rules cannot be used for a separate db. How do you implement the rules in your stack?
I'm using the Firebase API for authentication. Once authenticated, my backend will store the session as a cookie. I store the user ID in a postgres DB.
It's great because its free and I don't have to handle emails either for account creation or password reset etc.
Let me know if you have more questions or some help. Happy to send some resources your way.
@cao_wang
You can see the API in use here:
https://www.youtube.com/watch?v=esqNgnayVE8
I would use Rails as I am more productive in it
PostgreSQL / PostgREST + Node.js
I do like Rails but recently with the departure of many folks from the community things got a lot worst in terms of framework and package maintenance
Firebase / Supabase if I want to launch something really quick. But for my current project I needed a really complex back-end with functions depnending on each other.
So I decided to use AWS and couldn't be happier with it(except for the price 😂)
Personnaly I have a nextjs+tailwindcss+express/graphql boilerplate that I use in almost all my projects
Next.js + NestJS + Postgres
NestJS actually makes use of ExpressJS under the hood, so I vote for it
I prefer express js or nestjs. Recently I had to search for a new saas idea and here’s a note I’ve made https://whyboobo.com/webservice/nodejs-frameworks-comparison/
Am I the only one using Strapi ?
I've been loving Cloudflare Workers + Durable Objects. It's fairly new and non-traditional, and I had to custom-build some of the routing pieces, so it's not for non-developers (or for moving extremely fast). But it has the huge advantage of being built directly on an edge network, so latency is virtually non-existent. I'm going to be open-sourcing a routing library I built for the product, it's been a huge help in improving readability of the business logic layer.
https://feathersjs.com !
fabform.io -- We build out fabform.io all in javascript. Node.js in the backend.
I would say just use the language you already know.
Serverless AWS. API Gateway, Lambda, DynamoDB, etc. I don't want to worry about code that isn't business logic and I want the option to scale without worrying about how.
The reason I do that is because it gives me more flexibility and options in case I'd like to migrate my data to another vendor/provider. If that happens, I would not have to recode the whole authentication system in my app.
https://chudjen191.com/หวยลาว/
we've built the first version of https://acreom.com on django, was simple to learn, maintain and worked perfectly for the MVP.
It really depends. As a python developer myself, Django or Flask is probably my go to tech for the backend. However for personal projects, I might try out .net core (since I'm learning it). I guess your best bet will be to use whatever you're most comfortable using. As a sidenote, I'd also recommend taking a look at the pricing if you want to go the BaaS route.
I've been using Rails after going through a bit of a 'world tour' of frameworks for the better part of two years.
Mind you, I didn't pick Rails because "it's the best" (very subjective), but because I felt it offered the best mix of speed to MVP, maintenance, community and syntax for me.
At the end of the day, I think the advice that's offered of picking what you know is still the best.
I'm probably an odd one out here.
I use .NET Core for the backend and Vue 3 for the frontend.
It takes quite a bit of time to build out a fully fledged backend this way but I've been working with C# for a long time and I know my way around.
My template is a basic to-do list (Trello style) web app. But it's got all the bells and whistles of a fully fledged SaaS app (landing page, auth, stripe, email, etc). So when I get an idea for a new app I can spin up a new instance quickly and get cracking (or should I say hacking.)
The best part is I actually use the to-do list app myself to keep organised. So I'm always tweaking things in the template as I spot them.
I'm surprised that Ruby on Rails / Django isn't getting more mentions. Also JAM stacks (Javascript/APIs/Markup) is a solid choice.
Me personally I'm partial to Rails.
Although I have a background in Engineering I went with no code for my latest project. I built a prototype within just 2 days. I am using that to validate the idea, making tweaks and changes to evolve it into an actual MVP. I saved a lot of time compared to building it myself. I spend that time telling people the product exists. It may turn out people won't like it - so my approach me keep my risk. And if they do like it, I will happily build the thing for real.
I am using Bubble, Airtable, Sheets, Mailchimp
Going no code was not an easy decision. But I am super happy with it. Of course, depending on the type of project your mileage may vary.
https://appwrite.io/
I bootstrapped a Django project I use everytime I want to start a new project and that I keep upgrading with new features whenever I add something it didn't exist for the new product I'm building ...
which bootstrapped django app do you use?
something I've built brick by brick over time
have you open-sourced it? if yes then can you share it
I didn't not open source it yet but I did open source some of the libraries I developed for it :
https://github.com/rednaks/django-async-orm
https://github.com/rednaks/django-ip-geolocation
For a long time, I was using Laravel + ReactJs with InertiaJs to connect the frontend with the backend.
But I was running into context switching a lot so I moved to AdonisJs + Vue with InertiaJs and I find this combination awesome to work with mainly because it's all JavaScript/TypeScript now and Inertia is really great solution to connect Vue/React with Adonis without needing to write rest API just to connect the frontend with backend.
Adonis lacks some of the awesome features of Laravel but I'm happy with the trade-off because there is so little context switching now, also deploying node apps via pm2 is very easy compared to working with PHP-FPM.
For storage, I use hosted MySQL from Digitalocean because the database is something I don't want to self-manage.
Made a minimalist boilerplate for ExpressJS people so you can get your API up and running in 3 minutes.
It's free and public you can grab it here.
This is cool. I wish I had found this 4 weeks ago :-/
Thanks, you can star it for your next projects 😊
I'm using firebase on the back and expo. Pretty good experience so far! (First-time builder speaking here)
The sooner you step away from Firebase as your backend solution, the better. Writing all that "backend" code on your frontend? God forbid that you want to have your product on another platform and have to write all that Firebase spaghetti code again!
Great for analytics though.
You can use DAO pattern with Firebase. Realtime can be made to work using observer pattern.
I've been going for Next.js and Firebase. As a relative newbie to engineering, Firebase was the most intuitive for me, and it's project containers make much more sense in my head than what I've tried in AWS.
I am building https://shaperoute.com/ using postgraphile and https://www.manmeho.com/ using Hasura on the backend.
Both are graphql engines on top of Postgres database, so I get all the CRUD operations for free.
Rails
Azure Static Web App:
I've started experimenting with Xano which is a no code backend builder that auto-scales. So far, I've been pretty impressed with how quickly I was able to spin up an API to extend an existing self-hosted service into a potential SaaS offering.
Did you try Strapi? If yes, why you chose Xano over Strapi.
I haven't tried Strapi, but it has come across my desk. However I'm currently investigating no code solutions to see what is possible with low/no code solution.
I tend to experiment a lot. one of my favorites though is supabase. lately i gained some experience with craftcms and its kind of awesome to be honest
ApiPlatform (API framework based on Symfony, Doctrine and PHP)
I would use Clojure + Pathom + Postgres to build the backend services. It scales well and enables powerful abstractions without much code complexity. We are building Vade Studio that allows unifying multiple integrations using these underlying tools. It helped us to build a highly configurable backend system. We started with a standard API generation framework and quickly extended it to create an analytics pipeline based on customer requirements without any hassle.
I would hire a python/Django developer for the MVP
what are you building ? maybe I can help
sorry. actually, I was trying to say that if I had an option to develop the MVP, I would had choose Django.
I built the WBE Space with Hugo deployed in Netlify, Firebase Real Time DB and Google Functions...
Do you ever think of self-hosting it and moving out of vendor lock-in?
nope. it has not crossed my mind so far...
I usually go with MongoDB Realm.
AWS CFN (Cloud Formation Templates) and SAM (Serverless Application Model)
I recently built a new project with Firebase and it's been a pretty wonderful experience.
I'm slightly old school. I used to always be the dev that wanted to build things from the ground up (Mongo, Postgres, Express, etc). But honestly if you have fairly simple data requirements for your app - Firebase is pretty hard to beat and you can hit the ground running pretty quickly!
Now that Firebase has moved to Pay as you go plan. How does it compare with your previous approach (ground up) in terms of monthly expense?
So far it's actually more affordable in the beginning stages. The app isn't getting much traffic besides myself, however. Still under heavy development.
I run another app that's built on a traditional stack (Node/Mongo/Express) and while it's slightly more expensive to start, mainly due to the DB, the costs haven't really increased that much with more user adoption.
With Firebase's pay as you go plan, there will definitely be a breaking point where it will cost more eventually with increased usage.
Firebase (+ Tailwind)
Why do you prefer Tailwind over Bootstrap5?
Tailwind and its partners provide a set of preconfigured UIs from a form input to a landing page itself, which makes it easy to start a SaaS project.
Surprised no one mentioned Parse Server.
Next.js + Tailwind + Parse Server
I am surprised too considering that Parse has real time database capabilities with live queries.
Ruby on Rails
Ruby on Rails
asp.net (core). All the benefits and better code/performance/support in the long run.
I use JHipster and heroku.
Love to see other engineers using Firebase. It's so fast and can scale 'OKAY' into growth thanks to security rules.
Querying is the big con imo.
For me: Express w/ Bootstrap EJS templates (Bootstrap + Vanilla JS) w/ MariaDB backend deployed on Digital Ocean.
Once the project needs to scale I switch Frontend to React and look at the best Backend/Serverless options for what my project entails/provides best cost.
Rails via Jumpstart Rails
Xano
Flutter and Firebase.
Why does anyone use Laravel anymore?
Laravel is quite popular and you can get developers easily.
A lot of people are also using Laravel because of Spark.
Because they haven't discovered that Symfony has getparthenon.com? /s
But seriously, Laravel has lots of libraries, lots of cool stuff built on top of it, well maintained, and designed to build things quickly. I understand why people like using it. I also understand why people hate maintaining Laravel applications which is why I choose Symfony.
Could you please elaborate on why people hate maintaining Laravel applications?
It's designed to make building things easy. This means the initial creation part is easy and enjoyable. But because that part is easy it creates lots of dependencies and it's hard to swap out things.
Mines is getparthenon.com which is built on top of Symfony and PHP. It lets be build a MVP in about 12 hours or so. I'm also able to choose between MongoDB and Postgres for a database depending on what I'm building.
But I did build getparthenon.com to let me build things quickly so kinda makes sense it's my favourite.
I would be interested in knowing how much traction is it receiving from the community. In fact, that is the purpose of my original post and the poll.
From this community, I don't think much. Most of the sales have been from before I started being active in this community. But I'm hoping when they get to see how it's progressing more folk will want to save themselves time and energy.
Or maybe they like the cool voucher names I create like IGOTFIRED that is for 50% to celebrate going full-time on it. xD There is defo going to be more fun voucher names in the future.
Wtpsports.com - Laman slot mesin online gratis yang menyediakan tutorial permainan mesin slot komplit dan isu komplit dan komprehensif mengenai cara main mesin slot online dan beragam produk slot online terbaik.
When it comes to building a Minimum Viable Product (MVP), the backend is a crucial component. In fact, it's one of the most important parts of the whole process.
As a frontend developer, I typically don't have much input into what backend is chosen, but I do have some thoughts on what makes a good backend for an MVP.
First and foremost, the API should be well-documented and easy to use. If there are any questions about how to integrate with a third-party service or API, it should be possible to find answers within a few minutes' worth of searching. The API should also be stable—if you're trying out a new backend for your MVP and it suddenly breaks because of an update from their side, that's going to cause problems down the line when you need it working again (assuming you decide not abandon the project).
You also want something that's easy to set up and maintain—you don't want to spend valuable time learning about all the ins and outs of maintaining this system if you can help it!
Finally, make sure that whatever backends you choose will meet your needs now AND later on down the line as well.
This comment was deleted 17 days ago.