Back

The Journey

How I built a SaaS in less than 6hrs that got more than 2200+ active users and over 10k+ Subscritions processed



The Engineering

The tech stack that made it all

Next.js + TS ftw
...and a sprinkle of Vercel Functions
...and a dash of my own payment gateway
AI : OpenAI GPT-4o mini
DB : Supabase
hosted on vercel

The Payment Infrastructure

The Infrastructure that powered the whole operation, is on its own, a work of art. No payment gateway in India supported unregistered businesses, espicially unregistered SaaS businesses. The go-to solution, Stripe was not offering UPI. Given the demographics of the users, UPI was the only payment method that made sense... easy to use, no credit cards, no hassle. No one ever wanted to enter their card details on a website even if it was on Stripe's page (obvious kids)

...so that left me with no other choice

I built my own payment gateway

I wanted to receive payments directly to my bank account, and I also needed to verify if the user had paid or not. So I built a system that would get the payment screenshot from gpay, paytm or any other UPI app, and then verify the payment by checking the transaction ID and the amount.

A function will be called with the extracted data from the screenshot, which would verify the payment and activate the subscription for the user. A function that would retrieve my bank statement and cross verify the payment with the transaction ID and amount.

The system was designed to be secure, with multiple layers of encryption and authentication. The payment verification process was automated, ensuring that users could access the tool immediately after making a payment. The system followed an Optimistic Validation approach, where the user would be granted access immediately after making the payment, and the verification would be done in the background during high loads.

When you deliver genuine value, people dont mind paying for it. Money is not the problem, its the problem you solve.
I believe in humanity and the Optimistic Validation approach. Not giving even a single bad experience for my consumers than making sure they pay for it is my philosophy. Because I know that when you deliver genuine value, people are willing to pay for it. And the Optimistic Validation approach has also been favourable : there has been few instances where people thoght they beat the system only to realise that their subscription got suspended right when they want to actually use it.

Everytime a txn was made, i would get a webhook. And there were 2 types of webhooks I got
Webhook SuccessSuccessful Subscription
Webhook FailureSkid Attempt


This obviously made things easier for me to overview.

The Database

It was the most important part of the whole system. It monitored 2 things mainly.

1. user data

    phone, email, name, password, subscription status, remaining_tokens, total_tokens_used, subscription_expiry, connected_accounts (IPs),
    and other metadata

2. txn data

    phone, transaction id, amount paid, payment method, timestamp of the payment, sender
    and other metadata
The databases were designed to be simple so that all the data can be accessed easily and quickly. The user data was stored in a single table, with each row representing a user and their associated data. The transaction data was stored in a separate table, with each row representing a transaction and its associated data.

This meant it is easy for me to do any changes to any of the user accounts as all i had to do was filter the user by the phone number and i can change anything i want - all in a single place.

Supabase project dashboard was my admin panel

The Serverless Architecture

The complete system is serverless - from payment management to everything. All. Deployed. As Vercel Functions.

So its safe to say that it costed me absolutely NOTHING to run the whole operation.






The Marketting & Management

Marketting was done through word of mouth, and it spread like wildfire. Simple.
Build something that people want, and they will come to you.

...and thats not it all that was to the secret sause.
It was one fine morning sometime in August 2024, just 3 days after TTv2 was launched. I wake to bunch of texts saying the same thing - TT aint working mannnnnnn. I am confused, the probelem seemed to be with only the AI search, and check my OpenAI dashboard by chance...OverloadIt was the initial days of TTv2 and inorder to mass spread the word, I made it COMPLETELY FREE. And it was a huge success. The user base grew exponentially, got all the publiciy i needed that set me up for the next 8 months. The only problem, I didnt anticipate this big of an adoption, espicially not in just 3 days!

This was supposed to be my user aquisition cost and the money i spent on marketing. And this. This one screenshot went on to become the justification for switching to a paid model. And the narrative sold. Looking back, this was barly even fraction of the money I made on a monthly basis, and it was the best decision I made. The bill ended up being around $640 that month.

ps: i pulled a big brain move and eod didnt even end up paying for it. So that still keeps our running + investment costs at almost nil.


Sample marketing 101. done right


I aimed in providing the best user experience possible. Direct live support via Discord. Simplified instructions and video tutorials for almost everything.

Dispite all this, i realised that people are still dumb. They still ask the same questions over and over again, even after reading the instructions. When people pay money for something, they just switch off their brains afterwards.


The biggest asset for the whole marketting was - FOMO. The fear is what made people use TT, fear of missing out. Subscriptions were made in clusters - a friendgroup sees one or 2 people using it, and they all want to use it too. And then they all start using it, and then they all start telling their friends about it, and then their friends start using it, and so on. It was a chain reaction that led to the exponential growth of the user base.

Because everyone wanted others too to use it cause they felt scared to be the only one. those who didnt use it were scared that they will be the only ones who wont be using it and missing out.

Why

The Origin story how it all started...





It was more than just a SaaS tool, it was a phenomenon.
a phenomenon that that taught me more than I could ever imagine.