DEV Community: Praveen T

4 products, 1 broadcasting infrastructure, 3 weeks, 1 non-technical founder

Praveen T — Sat, 18 Apr 2026 08:05:40 +0000

This is not a flex post. This is what happens when you stop planning and start shipping.

I run NeuralChain Hub, a small company out of Dubai. Between March 27 and April 17, 2026, we shipped 4 digital products and built a production-grade live broadcasting relay server. None of this was planned as a batch. Each one came from a real problem that needed solving right now, not next quarter.

Here is what we built, why, and what I learned.

Product 1: The Guest (the-guest.org)

My daughter is 13. She needed a period tracker that wasn't designed for adults, wasn't creepy with data, and didn't show her fertility content she shouldn't be seeing at her age. Every app on the market failed at least one of these.

So we built The Guest together. She is the creator behind the product. My wife handles the education side. I handle the technical building.

The website is live at the-guest.org with real users. The app is built in Expo React Native with 62 features including cycle tracking, mood and symptom analytics, calendar, age-filtered recommendations, and full COPPA compliance. Under 13 users never see fertility or mature content. That was a non-negotiable rule from day one.

Stack: React for the website, Expo React Native for the app, Neon PostgreSQL in Frankfurt, NextAuth with verification codes (not Google OAuth, because minors). Both the website and app share the same database and API.

The hardest part was not the code. It was getting the medical logic right. Ovulation calculation must follow cycleLength minus 14. Period records only get created on day 1 of a new cycle. Predictions always use the profile's last period date as the primary source. Get any of these wrong and you're giving someone incorrect health information. That kept me up more nights than any bug.

Product 2: SportzGenesis Arena (sportzgenesisarena.com)

This one I wrote about last week on dev.to. Free sports prediction platform covering IPL, Premier League, Champions League, and World Cup 2026. No gambling, no money, just predictions and points.

58 users after the first week. Built with Next.js, Neon PostgreSQL, Vercel. Two different scoring systems for football and cricket. Match chat, private leagues, 16 badges, festival bonuses.

The interesting thing about building this alongside The Guest is how different the problems are. The Guest is a health product where accuracy is everything and one wrong calculation can hurt someone. SGA is a sports product where engagement is everything and nobody gets hurt if your leaderboard has a minor bug for an hour.

Same database provider, same hosting approach, completely different mindset required for each.

Product 3: The Guest App (Expo React Native)

Same product as the website but native mobile. 62 features, 99% parity with the web version. Onboarding flow, settings, dashboard, tracker, calendar, insights, all of it.

Building the same product for web and mobile simultaneously taught me something. Your API is the product. The frontend is just a view layer. Once the API was solid, the app was mostly about adapting the UI to mobile patterns. Swipe instead of click, bottom tabs instead of sidebar, pull to refresh instead of reload button.

The app is ready for App Store submission. That is a separate process involving Apple review, screenshots, descriptions, and the specific way Apple wants you to explain data collection to users.

Product 4: NeuralChain Hub Digital Products (neuralchainhub.gumroad.com)

While building the tech products, we also launched a Gumroad store with 4 digital products. AI Toolkit, a free lead magnet, a legal bundle for freelancers, and a freelancer starter kit. My wife is the public face of the brand. These are template and guide products for other entrepreneurs.

This was the fastest to ship because Gumroad handles everything. No code, no hosting, no auth. Upload PDF, set price, publish. The lesson here is that not everything needs to be a custom-built platform. Sometimes a marketplace product on an existing platform is the right answer.

The Broadcasting Infrastructure

This is the one that surprised me the most.

A client needed a production relay server for live cricket broadcasting. Venues in different countries stream via SRT to a central relay, which then pushes to multiple broadcast takers including major sports networks and streaming platforms. This is not a hobby project. This is live television infrastructure where downtime means millions of viewers see a black screen.

We set up a Hetzner cloud server in Germany running Nimble Streamer. SRT ingest from multiple venues on dedicated ports. Output via SRT push, RTMP republishing, and MPEGTS to different broadcast partners. Each venue gets its own port. Each taker gets its own output configuration. The server handles protocol conversion, so a venue can send SRT and a taker can receive RTMP without knowing or caring about the difference.

The port convention system was important for scaling. 9900 range for ingest, 10000 range for outputs. First two digits identify the venue, last two identify the taker. When you are managing multiple countries and multiple broadcast partners, this kind of structure prevents chaos.

Total infrastructure cost: about 87 dollars a month for a server that handles production broadcasting for international cricket.

The technical stack for this is completely different from everything else I built. No JavaScript, no React, no databases. Just Linux, streaming protocols, firewall rules, and a deep respect for what happens when a live feed drops during a match.

What ties all of this together

On paper these products have nothing in common. A period tracker, a sports prediction game, digital templates, and broadcasting infrastructure. But they share a few things:

They all solve real problems for real people I know personally. My daughter needed a tracker. My friends needed a prediction game. Freelancers in my network needed templates. A client needed broadcasting infrastructure.

They all run on boring, reliable technology. PostgreSQL, not MongoDB. Next.js, not the framework that launched last week. Ubuntu LTS, not the latest unstable release. Hetzner, not the most expensive cloud provider.

They all shipped imperfect and improved after launch. Every single one of these products had bugs on day one. The Guest had a cycle calculation error that took 3 iterations to fix. SGA had a match sort order bug that showed completed matches above upcoming ones. The broadcasting server needed firewall rule adjustments after the first test stream. None of that matters as long as you ship, listen to feedback, and fix things fast.

The non-technical founder reality

I keep saying I am not a developer because I think it matters for context. Everything described above was built with LLM assistance. I prompt, I read every line of generated code carefully, I test, I debug, I make architectural decisions. The LLM writes the syntax. I make the product decisions.

This is a new way of building things and I think it is underestimated. The bottleneck for most products is not the code. It is knowing what to build, for whom, and why. If you have clarity on those three questions, the code is increasingly solvable with tools.

That said, the broadcasting infrastructure humbled me. Streaming protocols, port management, codec settings, and latency tuning are not things you can vaguely prompt your way through. You need to understand what SRT does differently from RTMP, why keyframe intervals matter for segmented streaming, and what happens when packet loss hits 2% on a live feed. Some domains still require deep understanding even if you are using tools to execute.

Numbers as of April 17, 2026

The Guest: live website with active users, app ready for store submission
SportzGenesis Arena: 58 users, 215 predictions, 6 private leagues, 13 SEO pages
Gumroad store: 4 products live
Broadcasting relay: production server handling live international cricket feeds
Monthly infrastructure cost across everything: under 100 dollars

What is next

Getting SGA to 500 users before World Cup 2026 kicks off in June. Adding push notifications through PWA so existing users actually come back. Writing more SEO content so Google sends organic traffic. Eventually building a white label version of SGA for sports federations and colleges.

For The Guest, the immediate next step is App Store submission and then iterating based on what real users ask for.

For the broadcasting side, scaling the port and venue system to handle multiple concurrent tournaments across different countries.

Thanks for reading. Happy to answer questions about any of these in the comments.

How a Non-Technical Founder Built a Multi-Sport Prediction Platform (Lessons from Week 1)

Praveen T — Thu, 16 Apr 2026 11:43:44 +0000

So I shipped SportzGenesis Arena last week. Sports prediction platform, covers cricket and football, no gambling. 58 users right now. Here's what actually happened.

Why I built it

Every IPL season, my friends and I argue endlessly about match outcomes on WhatsApp. Same thing with Premier League. Nobody keeps score. Nobody remembers who called what. The arguments just reset every week.

I looked for an app that would let us track our predictions. Options were:

Dream11 and Fantasy Premier League. Too complex. Nobody in my group wants to research player lineups for 2 hours every weekend.

Superbru. Fine platform but looks like it was built in 2010. Clunky on mobile.

Most other fantasy apps. Either gambling-based (illegal in a lot of places) or too focused on one sport.

Also, here's a detail I didn't know until I started researching: India passed the Promotion and Regulation of Online Gaming Act in August 2025. Real money fantasy sports are effectively banned there now. Dream11 had to pivot its entire model. That left a gap for simple, free prediction games.

So I decided to build one.

I am not a developer

Need to be upfront about this. I've never taken a CS class. I can read code but I don't write it from scratch. What I do have is patience, an LLM subscription, and enough stubbornness to debug stuff at 2am until it works.

If you're a non-technical founder reading this thinking you need to learn to code before you build, you probably don't. You need to learn how to ask good questions, test things properly, and notice when something feels wrong. The rest of it, tooling handles.

The stack

Kept it boring on purpose.

Next.js with App Router, TypeScript, Neon Postgres (Frankfurt region), NextAuth with Google, Vercel for hosting. Plain CSS, no Tailwind. Total monthly cost right now is one dollar for the domain.

If you're bootstrapping, don't pick trendy tech. You don't have time to debug a bleeding-edge framework AND your actual product.

What the product does

Two sports, two different prediction formats.

For football, you predict the half-time and full-time scores. Max 8 points per match. Exact score gets 3, correct result gets 1, and if you nail both HT and FT exact you get a 2 point bonus.

For cricket, you predict five things: toss winner, toss decision, first innings score, match winner, and second innings wickets. Max 10 points. Cricket needed more prediction fields because a cricket match has more meaningful events than football.

Predictions lock 30 minutes before scheduled toss for cricket, and at kickoff for football. The 30 minute buffer was important because cricket fans react to team news and pitch reports that drop an hour before the match. If you lock predictions too early, you lose engagement. Lock them too late, people wait until the toss happens.

The cricket thing was harder

Football is clean. 90 minutes, one winner, done. Cricket has:

A toss, which is a separate event with its own outcome
Two innings with wildly different strategic considerations
Rain interruptions, super overs, DLS method calculations
Last minute lineup changes that matter a lot

I cut corners. Instead of modeling all of this perfectly, I shipped the prediction layer first and decided I'd deal with edge cases as they came up. A rain-shortened match still works with my scoring, it's just less points available. Good enough for v1.

Social features were the surprise

Thought people would come for predictions. They actually came for the chat.

I added a match chat feature almost as an afterthought. Each match gets its own comment thread where people can talk while watching, react with emojis, reply to each other. Most-used feature after predictions themselves.

Turns out people don't just want to predict. They want to predict loudly, in front of their friends, and then roast each other when predictions go wrong.

Also added private leagues. Users create a league, share the invite code, everyone in the league sees a separate mini-leaderboard. This is where most new signups come from now, people inviting their friends to compete.

Growth is the hard part

Product was the easy bit. Getting users is where I'm currently struggling.

What worked:

WhatsApp to friends and family. First 15 users, 100% conversion, immediate feedback. No surprise here.

SEO landing pages. Wrote 13 pages targeting specific keywords. Too early to measure impact, but Google indexed them within 5 days. This is a slow compound play, probably 6-8 weeks before I see real traffic.

Cultural moments. Added a small joining bonus during Vishu (Kerala new year) which brought in a handful of users from that community. Worth doing for any festival your target audience cares about.

What absolutely did not work:

Instagram stories. Zero impressions. Found out stories only reach existing followers. When you have 10 followers, you get 10 views. Complete waste of time.

Twitter. Brand new account, algorithm suppresses you for weeks. Posted content for two weeks straight, got nothing.

LinkedIn company page. Reach is apparently 10% of personal profile reach. Should have posted from my personal profile instead.

Meta ads. Account got banned mid-campaign. Appeal rejected without explanation. Don't know what happened.

WhatsApp status updates. Worked initially, then my friends got tired of it. Had to stop because it was becoming spam.

The uncomfortable truth

Indie hacker Twitter makes growth sound easy. "Just ship it." "Build in public." "Growth hack to 10k users."

Reality for a small account with no following: social media platforms actively throttle your reach. Paid ads are gatekept by policies that change without notice. Organic reach on a new account is basically zero.

What actually works is slower. Direct messages to real people. SEO that compounds over months not days. Making something genuinely good so that the users you do have tell their friends.

Lessons I'm taking into week 2

Build for real users not imaginary ones. My first version had an AI generated match preview feature. I spent days on it. Nobody used it. Deleted it.

Simple beats clever. Predict the score is something casual fans understand. Building a 15 player fantasy squad is not.

Community features punch above their weight. Match chat took almost no engineering but drives more engagement than sophisticated features I built.

Growth is a separate full time job. Building is fun and creative. Growth is grinding. Writing copy, sending messages, replying to comments. Most founders prefer building. Most products fail to find users. These two facts are connected.

SEO should start day one. I added landing pages in week two. Should have been week one. Google takes weeks to index and rank. Every day without SEO pages is lost compounding.

B2C proof matters before B2B. I have plans for a white label version for sports federations and colleges. Nobody will sign up with a platform that has 58 users though. The consumer traction is the credibility for the business pitch.

What's next

Get to 500 users. That's the number I need before I can credibly pitch to federations, colleges, corporate sports organizers.

Keep writing SEO pages. Google rewards consistency.

Launch on Product Hunt after I hit 500 users. Not before. A Product Hunt launch with 58 users is embarrassing. With 500 users, it's a launch moment.

Probably write a week 2 update here on dev.to if I have anything interesting to say.

Feedback welcome, especially roasts of my homepage. Growth tactics that worked for you, please drop in the comments.

www.sportzgenesisarena.com

Thanks for reading.