Ask HN: What rapid development framework do you use for your MVPs?
20 comments
Pharo Smalltalk. For web apps there is Seaside and small restful stuff like Teapot. But I use plain Zinc. Anyways, the famous Smalltalk interactive development makes for the fastest and intuitive experience ever (for me at least). Dolphin Smalltalk for Windows-centric desktop development. And always with SQLite3 at the backend.
Second would be Lua. For web it would be Lua Lapis, which runs in OpenResty. I have also been looking at Lua Tarantool for generic servers.
Second would be Lua. For web it would be Lua Lapis, which runs in OpenResty. I have also been looking at Lua Tarantool for generic servers.
Best tool is the one you already know, isn't it? Rails for me as well. Is there something that you feel is missing from Rails for rapid prototyping?
As an aside (and no offence intended) I feel like proliferation of frameworks for all languages and all platforms is just folks reinventing the wheel. If I was a pythonista or a javascript guy, I'd certainly do something like Django or Node. ¯\_(ツ)_/¯
As an aside (and no offence intended) I feel like proliferation of frameworks for all languages and all platforms is just folks reinventing the wheel. If I was a pythonista or a javascript guy, I'd certainly do something like Django or Node. ¯\_(ツ)_/¯
Agree with the others: use whatever is fastest for you. To me that is create react app, Chakra UI Pro, and Flask / SQLAlchemy.
I'm in the process of creating my own boilerplate project setup with this, so the next time around it will contain all the basics (auth, billing, etc.).
I think divjoy is a productized version of that, but I've never used it. Heard good things though!
I'm in the process of creating my own boilerplate project setup with this, so the next time around it will contain all the basics (auth, billing, etc.).
I think divjoy is a productized version of that, but I've never used it. Heard good things though!
Mostly firebase, used django in the past but leaning towards flask lately (mostly because it's fast to put it on app engine).
We also started Saasform [1] to automate website + auth + payments. It's designed as a microservice so it can work easily with any stack. Still in beta, but if anyone takes a look I'd be more than happy to hear what you think.
[1] https://github.com/saasform/saasform
We also started Saasform [1] to automate website + auth + payments. It's designed as a microservice so it can work easily with any stack. Still in beta, but if anyone takes a look I'd be more than happy to hear what you think.
[1] https://github.com/saasform/saasform
Although it might not be perfectly positioned to build a consumer facing web-app, we wanted to make it simple to build web apps. So we built https://lowdefy.com - It makes building web apps as simple as writing a query.
I would like to challenge anyone to show me a faster way to build out a web app of medium complexity, especially one that is more focused on working with data.
If it is only demonstrating and testing concepts that you are after, Lowdefy can be a nice quick win.
If you are looking for a more polished consumer experience in your MVP, I would consider Flutter, or Next.js, even Meteor.js with React.
If you are willing to consider the No-Code route, you can also build fancy things with Bubble.io or Webflow.
I would like to challenge anyone to show me a faster way to build out a web app of medium complexity, especially one that is more focused on working with data.
If it is only demonstrating and testing concepts that you are after, Lowdefy can be a nice quick win.
If you are looking for a more polished consumer experience in your MVP, I would consider Flutter, or Next.js, even Meteor.js with React.
If you are willing to consider the No-Code route, you can also build fancy things with Bubble.io or Webflow.
Firebase + svelte.
Firebase gives you auth, serverless database that you access from the client directly, storage etc.
Svelte gives you low overhead, easy to write and easy state management components. it's very close to plain HTML & JS.
Firebase gives you auth, serverless database that you access from the client directly, storage etc.
Svelte gives you low overhead, easy to write and easy state management components. it's very close to plain HTML & JS.
afterwards you are locked into Firebase.
If your MVP succeeds, you should probably throw away the majority of your architecture anyway, so you'll have the debt to address in the future regardless. Why not get up and running quickly?
I really have been liking the look of Svelte, and used it briefly for a toy project. Have you had success launching anything with it? It still seems a bit immature for my tastes. The last I had heard was that they basically rewrote everything in the most recent major version and that sort of turned me off. Might be totally wrong on that, though.
Yes, I have launched a couple of svelte (sapper) over the last two years. No regrets so far.
What they are rewriting is sapper, the next-like framework. It's probably subjective, but I think the ergonomics and low code/noise ratio worth the immature state.
Also my team loves it, JR engineers find it easy to learn, even our designer likes to contribute now that we are using svelte.
What they are rewriting is sapper, the next-like framework. It's probably subjective, but I think the ergonomics and low code/noise ratio worth the immature state.
Also my team loves it, JR engineers find it easy to learn, even our designer likes to contribute now that we are using svelte.
Another vote for Svelte from me. But my set up is:
Svelte with Sapper for the front end. It calls API endpoints in a Java Spring backend. I already know Spring very well so this setup is pretty fast for me.
Sometimes I use PostegreSQL but at the moment I am using MySQL.
I don't plan on using this stack for just the MVP though. If any one of my projects picks up traction I plan to keep this stack.
Svelte with Sapper for the front end. It calls API endpoints in a Java Spring backend. I already know Spring very well so this setup is pretty fast for me.
Sometimes I use PostegreSQL but at the moment I am using MySQL.
I don't plan on using this stack for just the MVP though. If any one of my projects picks up traction I plan to keep this stack.
Built my own, Restful - https://restfuldata.com
Takes a spec file similar to other SDL files and generates a containerized, structured Node.js API with a few niceties like a JS client library for use in the browser.
I use it to spin up my projects quickly. Latest one being https://VBGart.com, which I’ve spent under half a day on so far.
Takes a spec file similar to other SDL files and generates a containerized, structured Node.js API with a few niceties like a JS client library for use in the browser.
I use it to spin up my projects quickly. Latest one being https://VBGart.com, which I’ve spent under half a day on so far.
Scala, postgres, scala.js. I can work very rapidly with my tech stack but the entry barier is nôt low...
I did create recently https://insightrend.com/?page=term-trending
Wasnt too complicated for mé. A Google trends alternatíve. Gl.
Wasnt too complicated for mé. A Google trends alternatíve. Gl.
You can go quite a way with Wordpress. But I don’t like PHP programming so for backend I’d set up a nodejs server for anything I can’t achieve with WP plugins
Firebase + ChakraUI + React.js has been pretty solid for me lately.
That said, I used Rails for the first time for a small client last week and it was fairly easy process.
That said, I used Rails for the first time for a small client last week and it was fairly easy process.
You can look into laravel and maybe laravel spark if you're building a saas.
Lazarus / Free Pascal for something that runs locally on my machine.
Python for other stuff.
Python for other stuff.
Svelte / Routify + Hasura are so good it feels like cheating sometimes.
Long post, but I hope it helps someone out since it covers web,mobile, etc. MVP to me means being able to build something quick and iterate upon, so while it could be a throwaway project each time it still makes sense to have a preferred solution for each problem space.
I stick to tech I know best and also languages which offers a vast variety of available SDKs and dependencies to consume so I don't need to waste time re-inventing the wheel. I also try to plan for my requirements up front so I can best prepare. In this case, I am building an application that requires authentication, notifications, private + group chat, location awareness, and an admin sites for the CMS, partner portal, etc.
- For data storage, I use postgres for persistence, and redis for caching and other temp things. I'm not an expert at either of these, but I have used both extensively and have enough experience/interest to work through any issues that may come up.
- I use nodeJS + express to write backend services. Example: My web and mobile clients consume a REST service that interacts internally with a data layer service. Since the code is organized very simply and modular, it makes it easy for me to spin up new services that consume existing services, or makes it easy to extend. I have a real time service that could use the same data layer service if needed.
- I try to use SDKs that can solve my problem cheap + quick. This is where planning up front is also valuable because if you can identify which SDKs are available for the languages you are using, you will save lots of time and pain. For example, I'm using connectycube (free up to 10k users) for video + chat because they have flutter SDKs it takes literally 10 minutes to set up and be most of the way there. I use firebase for authentication (its free) on the clients, and for my REST service token verification. Firebase can also handle push notifications, so keep that in mind! I use mapbox for geocoding because it is cheaper than google maps.
- For infrastructure, I was using AWS but switched to DO managed postgres and droplets. I'm still using S3 and cloudfront.
The reason I did this is because I don't have much infrastructure experience and made a few misconfiguration mistakes on AWS that bit me. DO makes it dead simple for now so I can focus on just building my product and maybe some day I will be lucky enough to work with a team who knows AWS :)
- For web, I have the website (hosted using carrd, dead simple website builder. They have a cheap paid plan that lets me do things like add a contact us widget and add a google analytics), while the CMS and partner sites are being built using create-react-app. There is no consumer web app (mobile app only), so really those users will just see the site and go download the apps. Partners can access the partner portal to do things like adding their business listing and any promotions, etc.
- Mobile apps are built using flutter. I've dabbled with swift/swiftUI, kotlin, react-native, and flutter and today I prefer flutter for complex projects that I need on both iOS and android. Regardless of what path you choose, some tips I can offer when you're in the middle ground between a "this is more complicated than what's in the /examples folder" and "any top app that you use on a regular basis": • Spend a weekend going through the platform specific design guidelines. Open up your favorite top apps on iOS and android and use the same features on both. Notice the subtle differences. • Read the app store guidelines for both so you know what to expect. • Plan your application structure. What are the shared states, and how will you share data? Do you need to persist data on the device? What device permissions do you need to manage? What are the top level views? Do you need to make network calls? I used to never do this because I figured hey, its just a toy project!! But, as soon as that toy project turns into a passion pit, I need to make that new features can be added very fast instead of requiring a re-write. Just note the different moving pieces, and organize your code in a way that makes sense. • Nothing is perfect - flutter web is kinda meh and honestly I think its best anyway to just have separate mobile code, and web code.
Learned a ton over the pandemic, hoping to put all of this to good use and keep learning. All the best!
I stick to tech I know best and also languages which offers a vast variety of available SDKs and dependencies to consume so I don't need to waste time re-inventing the wheel. I also try to plan for my requirements up front so I can best prepare. In this case, I am building an application that requires authentication, notifications, private + group chat, location awareness, and an admin sites for the CMS, partner portal, etc.
- For data storage, I use postgres for persistence, and redis for caching and other temp things. I'm not an expert at either of these, but I have used both extensively and have enough experience/interest to work through any issues that may come up.
- I use nodeJS + express to write backend services. Example: My web and mobile clients consume a REST service that interacts internally with a data layer service. Since the code is organized very simply and modular, it makes it easy for me to spin up new services that consume existing services, or makes it easy to extend. I have a real time service that could use the same data layer service if needed.
- I try to use SDKs that can solve my problem cheap + quick. This is where planning up front is also valuable because if you can identify which SDKs are available for the languages you are using, you will save lots of time and pain. For example, I'm using connectycube (free up to 10k users) for video + chat because they have flutter SDKs it takes literally 10 minutes to set up and be most of the way there. I use firebase for authentication (its free) on the clients, and for my REST service token verification. Firebase can also handle push notifications, so keep that in mind! I use mapbox for geocoding because it is cheaper than google maps.
- For infrastructure, I was using AWS but switched to DO managed postgres and droplets. I'm still using S3 and cloudfront.
The reason I did this is because I don't have much infrastructure experience and made a few misconfiguration mistakes on AWS that bit me. DO makes it dead simple for now so I can focus on just building my product and maybe some day I will be lucky enough to work with a team who knows AWS :)
- For web, I have the website (hosted using carrd, dead simple website builder. They have a cheap paid plan that lets me do things like add a contact us widget and add a google analytics), while the CMS and partner sites are being built using create-react-app. There is no consumer web app (mobile app only), so really those users will just see the site and go download the apps. Partners can access the partner portal to do things like adding their business listing and any promotions, etc.
- Mobile apps are built using flutter. I've dabbled with swift/swiftUI, kotlin, react-native, and flutter and today I prefer flutter for complex projects that I need on both iOS and android. Regardless of what path you choose, some tips I can offer when you're in the middle ground between a "this is more complicated than what's in the /examples folder" and "any top app that you use on a regular basis": • Spend a weekend going through the platform specific design guidelines. Open up your favorite top apps on iOS and android and use the same features on both. Notice the subtle differences. • Read the app store guidelines for both so you know what to expect. • Plan your application structure. What are the shared states, and how will you share data? Do you need to persist data on the device? What device permissions do you need to manage? What are the top level views? Do you need to make network calls? I used to never do this because I figured hey, its just a toy project!! But, as soon as that toy project turns into a passion pit, I need to make that new features can be added very fast instead of requiring a re-write. Just note the different moving pieces, and organize your code in a way that makes sense. • Nothing is perfect - flutter web is kinda meh and honestly I think its best anyway to just have separate mobile code, and web code.
Learned a ton over the pandemic, hoping to put all of this to good use and keep learning. All the best!
I'm curious if anybody uses no-code platform, like bubble?
Thank you all!