While AI tools are great for boosting productivity, this post focuses on the foundational tech stack every developer can use to build and deploy a production-ready app from scratch. This tech stack is minimal yet extremely powerful, helping you and your team go from idea to MVP in just a few days while keeping costs low.
1. Node.js + Express.js - (Vratix)
A solid backend API is the foundation of a good project. Build a robust backend service early on, and you and your users will be happy for a long time.
However, setting up a new Node.js API can be daunting, especially when wrong decisions early on lead to massive technical debt. Setting up authentication and remembering how to handle JWTs and refresh tokens properly from your last project two years ago can also slow you down and introduce security vulnerabilities.
I use Vratix and their open-source modules to avoid all of that. The tool comes with a CLI that sets up a TypeScript Node.js project, follows the latest best practices, and lets you install their API modules super easily. I don't have to read Node.js setup guides anymore!
The tool is open-source and free to use.
Project link: Vratix
2. Next.js - (shadcn/ui)
React has been my favorite for years when it comes to building beautiful web apps. I have recently started using Next.js, an open-source web development framework for building React-based web applications. It comes with many optimizations, such as Server Side Rendering (SSR), dynamic and static props, image optimization using next/image, and much more (especially when deployed on Vercel—more on this below).
Creating buttons, inputs, containers, and other basic UI components is time-consuming, and they are often the same across various projects. That’s why I use a component library called shadcn/ui, which offers a wide range of beautiful UI components you can integrate into most modern frontend frameworks.
Next.js and shadcn/ui are open-source and free to use. I highly recommend using them together to get a working web app in a few days and avoid setting up boring UI components.
Project links: shadcn/ui, Next.js
3. Database + Storage - (Supabase)
Next, we need to take care of storing user data and any files our app might use. To do this, I use Supabase, an open-source Firebase alternative based on PostgreSQL. You can have a fully working PostgreSQL database in minutes with just a few clicks. Supabase allows you to query using their Data API or connect directly to the database using a connection pool package like pg (if you use Vratix, this is already set up - you just need to add the correct connection string).
The coolest thing about Supabase is the ecosystem it provides for seamless integration with all of its other services. Because all of their tools are connected to your PostgreSQL database, you can get S3 bucket storage with Row Level Access with minimal setup. If you’ve ever tried to secure an S3 bucket and get AWS permissions right on the first try, you know the pain.
Supabase is open-source, but it also offers a managed solution with a great free tier.
Project link: Supabase
4. Deployments - (Vercel & AWS)
For deployments, I use Vercel for the frontend and AWS for the backend.
- Vercel: Perfect for Next.js apps, with automatic deployments, global edge network, and free HTTPS. It offers loads of optimisations for your Next.js apps, caching, performance analytics. You can deploy a working web app in minutes without setting up SSL, static files, hosting etc.
- AWS: For the backend, AWS offers great flexibility. I usually deploy my services as Docker containers on EC2, for more complex backend services I use Kubernetes (AWS EKS).
This combination gives you a fast frontend and a scalable backend without costing a lot.
5. Serverless Functions - (AWS Lambda)
This one is a bonus, as I know some of you might want to set up serverless functions to handle specific logic, e.g., webhook handlers or processing files and data.
I’ve looked at all the available options, including Vercel’s and Supabase’s Edge Functions. However, unless you need your functions to execute super close to your users, AWS Lambda will work just fine. AWS has significantly improved the developer experience for setting up new Lambda functions over the past few months.