What SaaS boilerplate do you use?
What boilerplate do you use for things like templating and more common functionality of your app?
Trending on Indie Hackers
What boilerplate do you use for things like templating and more common functionality of your app?
I built my own framework / platform.
It's built on top of Symfony Framework ( PHP ) for backend and Quasar Framework ( Vue ) on frontend.
I use latest API technology called GraphQL. Even I have built lot of reusable components such as billing, user management, role and permission, Job schedules , queues etc. It also has the multi tenent architecture set-up.
Good part of my platform is I have also developed a code generator along with it. It generates both Backend and Frontend codes.
Hence, I can do a CRUD operations from API to UI in just 5 minutes. Everything working along with validation.
Interesting to see many people building their own. In the end I went that way as well but did share it for others to use.
The problem with many I have is that they are just too basic. Some mid-level dev with no real industry experience stitches together a few published pieces of tech and packages and sells them to generate a profit.
The founder wanting to make a real business then spends tens of thousands or hundreds of thousands developing common SaaS systems that every other SaaS company needs. Furthermore, because it's simply stitched together, it becomes really hard to develop since you are integrating many different development perspectives together.
The one I work on is built on top of Rails. It's interesting that people feel they need Javascript these days. It just makes things complicated and slow despite their claims. To get speedy Javascript will take way more labor than something like laravel or Rails. Labor translates to increased costs for you.
I really do like the Rails 7 and hotwire approach. I started mine before that was released so the architecture is aligned but not using the standard in some parts. It's blazing fast. Browse my app to see a near native web experience running on crap Heroku hardware. The average response is about 30ms. When it takes half a second I get annoyed.
From there I built in a decade of business process knowledge over top of that stack. Gdpr compliance, encryption, 2FA, tracking and analytics, billing systems for real companies with accountants, basic support systems, style framework etc. The end result, a tight core of effective business processes that I can clone and build high quality SaaS apps.
This is completely subjective. People should use whichever stack they are comfortable with. Someone with 10 years Javascript experience isn't going to be faster working in Rails or PHP.
I didn't say you shouldn't use javascript. It's just more effort to build the same project. Sometimes that is good. Sometimes it is not. For most founders scraping by, it is not. Money and time are valuable so why go with a dev stack that is slower to develop on apples to apples. Developers are not the only department of a company and catering to them too much is quite dangerous for founders.
If I build a javascript app with any form of state management on the front end, then you need to start building things like checking if the backend is offline, duplicate front end and back end error handling, and the list goes on. The cost to build a javascript app is often three to five times a backend rendered server side app and often ends up lower quality due to increased dev skill required to build it and deploy it.
You encounter an impossible distributed state problem that is most likely going to cause some sort of baseline of support bugs that increase costs for that department and lots of time tracking down obscure bugs. To top it off, javascript developers often lean towards the junior end and lack any sort of real-world architecture experience. (it's so easy to get started!) they end up building apps that are slow and unwieldy. It definitely doesn't help that there are no clear standards or conventions on how to build apps so many JS devs never get to learn what actually works well in companies. At least most mature frameworks give some structure and guidance to varying degrees (Ruby/Php a light structure vs Java / compiled typically have lots of structure from the start).
Remember, it is "Computer Science" and not "Computer do what I like". So no, I don't agree it's subjective. You can measure and track it back to costs effort. But I do agree it's up to the company to decide if they want to go with a more expensive stack to develop on. There are many valid reasons to do so. Crud-based apps don't really have a reason aside from wasting money or personal motivations/learning.
Lastly, you're right that a JS dev with experience is faster in JS. But that's a non-sequitur. If you're starting from scratch then you get to pick what is best for the business.
I am not sure your analysis is entirely objective. First of all, it seems that you are assuming that JS is equivalent to client-side programming. But this is not really true, you can not only use many other languages for client-side development, but you can also develop your server-side code in javascript (or more commonly these days Typescript) with a mature framework like NodeJS.
Secondly, from a computer science standpoint, developing the presentation layer of the app server-side or client-side does not make any difference complexity-wise, i.e. filling a template server-side vs getting JSON data client-side and filling that same template is the same amount of computation.
Now, the advantage of building the app client-side is that you can create much richer applications, it is just impossible to implement things like for example "drag and drop" without client code. If your app is just displaying a list then sure, no big difference. Probably the case is that most modern "javascript" apps are simply much more complex functionality-wise than the average plain rails app, which also explains the higher development costs.
I know almost nothing about this, but seems like most people is "building their own". Does it make sense? How is that more eficient than using something that's already out there and being tested and improved by a community?
@sachingk @kylegawley @strzibnyj
Most of the boilerplate are too basic. Few of them don't have proper modular design , permission system etc.
Also most of them are being developed by a person, who no more actively supporting it.
That's about risk mitigation.
On the efficiency side,
I see that most boilerplate don't have support for multi tenent architecture that way an application needs.
For example, I need a Slack like multi tenent architecture. Where Organization has workspace and workspace has channels etc.
For those reasons, it's worth building your own.
One more reason why did it because I want to use GraphQl for API and some Frontend solution that has capability to develop any kind of apps on any platform.
Which boilerplates have you used that are too basic and not actively supported?
I haven't used any. But I reasearch few of them in PHP and VueJs ..
Didn't find them very promising... That's why I built my own platform.
Bootstrap + rails
When I start building a new project, I use the Flask web application framework as my boilerplate. It provides a lot of core functionality that's helpful for getting a project off the ground quickly such as templating, common functionality (forms, local storage, etc.), and routing. It also makes it easy to integrate other tools that can be useful in the development process (database integration, authentication, etc.).
Have a look at https://www.bappaz.com, a visul app builder, start with data model, get instant ready made UI, customize with drag-n-drop editor, test out the app and then export full stack code base with React for front end, NodeJS/ExpressJS for backend. You can use it as is or as a starting point and customize as needed.
I also built my own, I find that learning and maintaining a "boilerplate" is usually more work than simply building my own, which in general works well enough considering that I am the only one that has to understand the code.
I built my own, although it is taking years. In hindsight i should have used this django react option is https://www.saaspegasus.com/.
I am also building mine: https://www.hypersaas.dev/
It is a Django, React/Typescript boilerplate.
Also built my own based on Flask: https://github.com/nuvic/flask_for_startups
Wrote a book about it: https://nuvic.gumroad.com/l/flaskforstartups
I use a combination of these two:
https://github.com/Blazity/next-saas-starter
https://github.com/vercel/nextjs-subscription-payments
Tech stack is: Nextjs, supabase, stripe, sendgrid, mailchimp
I built my own and then sold it to 365 customers → https://usegravity.app
Im curious , How much time it took you to build it ?
btw great work!!
Thank you, I have been working on it for over two years now.
I am building https://businessclasskit.com/ for myself. It's optimized for people that want to have a SSR monolith + API for customers (rather than SPA). I am including the tedious things that I believe apps should have like email confirmation + email change and reconfirmation + two factor auth + payments, etc. A big focus is on tests as well (the video on the website runs Rails system tests).
Why doing my own kit?
I want to be Rails-native - keeping all Rails default choices if I can. Most people deviate from these, but I actually like them!
Firebase + Any frontend frameworks
Visual Basic all the way
This comment was deleted 3 years ago.