Qurious

The Backstory

After I left my job, I took a year long break from programming. For the first time in years, I didn't touch a single line of code. Then by the end of 2024, the itch to write code again returned. This time I wanted to create something that would fill in all the existing gaps in my skillset. Something to make me a complete, well-rounded developer.

I decided to build a fully functioning SaaS product from scratch. It would allow me to learn a range of technical skills as well as understand the business side too - payments, marketing, growth, etc.

Picking the Idea

But before I even begin, what do I build? The possibilities were endless. Something that lots of people will use? Or something that a tiny specific but loyal minority will use? And on what topic? Fitness? Yeah that's something I like. But that's too generic and boring.

Hmm, how about Science? Yeah, that's something I also like. Then began my search into science-related SaaS tools and products. I found a few, and noted down their list of features and then slowly started making a list of requirements for my MVP.

That's how the basic idea formed. An AI powered search engine for scientific papers. Neat! Something that gets me excited and passionate!

Picking the Tech Stack

I already knew that since I was building an entire SaaS by myself, I'd need something that I wouldn't have to fight configuring, managing, and using every step of the way. I would've loved to use AWS and configure every small detail by hand (or rather by code), but I like actually shipping more. The best option was to then choose a BaaS (Backend-as-a-Service), as it would allow me to get the most code out the quickest. After searching and comparing a lot of options, I decided to go with Convex as it seemed to provide a good DX out of the box.

And after having written the first version of Qurious, I can say that was the right call!

Picking the Payment System

Now this is where my real nightmare began. Stripe had suspended its operations in India. So I started looking for alternatives. I picked Paddle. Spent a few days learning it and setting it up. Only to discover you can't even use Paddle until you already have money coming in! Alright then, I did some more digging, and found LemonSqueezy. It seemed promising, but alas, it got acquired by Stripe. Finally, I found a post by the CEO of DodoPayments on a subreddit for Indian developers, and they literally described the exact problem I was facing at the moment, word for word. So I decided to go with their product.

And with that, began the actual development of Qurious.

The Switch to a Monorepo

After two months of development, I was noticing a pattern: the application was basically becoming split into multiple different sub-apps, for example, the landing page and the actual app were distinct apps that could be separated out and the code could be made a lot cleaner. So I made the decision to switch to a monorepo.

It was my first time working with a monorepo. It was difficult to port the app to a monorepo at first, but after two weeks of hair pulling and dependency chaos I got it working.

And once everything worked, the payoff was immense. The separation of apps and packages, the code reusability. It gave me dopamine just looking at the folder structure.

The First Apps

After the monorepo was stable, everything felt easier to manage and separate. Unified design systems, backends, isolated frontends, separate marketing.

Then after months of hard development, I had the app ready to be used by others. Well, not just one app, but apps.

In the end, I had built an ecosystem of multiple web apps - www, app, blog, help, status, docs, api. All powered by a single Convex backend and unified design system.

Built with: Turborepo, Next.js, Convex, Vercel AI SDK (Gemini), Semantic Scholar, Clerk, PostHog, DodoPayments, ShadCN, TailwindCSS.

Qurious - A search engine with folders, papers, chats, notes, sharing.

Future Plans & Marketing

Qurious has been built after months of research, coding, and designing, but this is just the start. If making it is 30%, then marketing it is the remaining 70%. And that's the hard part.

My future plans are to market the SaaS. I'm learning marketing and the next step is to share it in subreddits, forums, and communities, etc. I've already shared it with friends and family and so far the feedback has been positive.