₦airaland Forum

I know, it will sound very unbelievable but friends, I truly really did built a very beautiful, scalable and sophisticated AI Predictions web app that's very responsive without a single line of code.
It started as a curiosity to build and design a personal machine that can scan football matches everyday, analyze them, score them and then filter out the best by ranking with confidence scores that's based on so many metrics.

The tough begining:
I was very skeptical at first, like, hmm, will this AI thing build another AI powered thing?
Well, it sure did, only that it has to be refined and perfected with the use of precise prompting.

It should have taken me around 1 week to complete the MVP and ship it to production, but I wasn't always satisfied with what I have achieve within that kind of short time, so I refined and refined and refined, and 7 days turned two weeks, two weeks turned 1 month and a month turned 8 months.
But eventually, I got something beautiful only the best of the best software developers can achieve.

Look at it...isn't it beautiful?!
It's still in it's early stage, let's say beta.
In few weeks to come, I'll have added so many brilliant features to help all sports fans around the world understand the depth of the game better❤️

If you have any remarks, words of encouragement, criticism or questions regarding this, you are free to let me know.

And oh, since the web app has moved from an idea to a real thing, with around 890 daily users, then If anyone is interested in partnering or investing in it's marketing and growth, please reach out to heyzilobet@gmail.com

Wow! you really try I doubt you used AI look like a project built by 5 developers to be honest . I found out u use React js and Supabase I don't know the api u used actually the hosting was Namecheap. You could sell it to acquire.com for millions of dollars or hundreds of thousands of dollars Or just looking for a wealthy investor to push it global.

Well, Great!
Couple of questions:
1. Did you build the frontend and the backend with AI?
2. did AI build the database for you or you are using third-party APIs?
3. Did AI do the testing for you?
4. Is the app in production now?
Answer these ones first, I have more questions

Alphabyte3:
Wow! you really try I doubt you used AI look like a project built by 5 developers to be honest . I found out u use React js and Supabase I don't know the api u used actually the hosting was Namecheap. You could sell it to acquire.com for millions of dollars or hundreds of thousands of dollars Or just looking for a wealthy investor to push it global.

Thank you so much for the uplifting remarks.
I actually built it alone with an AI and absolutely wrote no codes.
And as for the selling to acquire.com, I actually have considered that but as a newbie to website flipping, I do not really know how that works but I'll look into it from now to understand it better.

platymus:
Well, Great!
Couple of questions:
1. Did you build the frontend and the backend with AI?
2. did AI build the database for you or you are using third-party APIs?
3. Did AI do the testing for you?
4. Is the app in production now?
Answer these ones first, I have more questions

Hi, thank you for the input and I'll answer all of your questions accordingly.

1. Yes, both the front end, back end, contents on the blog page were all built using one single AI.
2. The AI built the database, including edge functions, crons, etc. It handled not just react or next.js gracefully, but also handles backend logics as well through supabase or even firebase.
3.As for the testing, the AI can test on site functions and logics to see if they're indeed working as expected and are not failing silently, but for load testing it & stress testing, I employed external services for that.
4. Yes, the app is in production mode right now and it's performing very well.

platymus:
Well, Great!
Couple of questions:
1. Did you build the frontend and the backend with AI?
2. did AI build the database for you or you are using third-party APIs?
3. Did AI do the testing for you?
4. Is the app in production now?
Answer these ones first, I have more questions

Using a Cronjob to fetch data from third-party like Sportradar or Lsports APIs can automate the process, making it more efficient. If you create your own APIs manually, it'll require daily updates, which can be time-consuming.
Before we proceed, do you have API keys for Sportradar or Lsport. There are Many alternatives.

Alphabyte3:
Using a Cronjob to fetch data from third-party like Sportradar or Lsports APIs can automate the process, making it more efficient. If you create your own APIs manually, it'll require daily updates, which can be time-consuming.
Before we proceed, do you have API keys for Sportradar or Lsport. There are Many alternatives.

Hi again, thanks for the subtle insight.
But actually, the web app is already integrated with sports data API provider; www.api-football.com
The web app currently uses this service to fetch and update match data and statistics, which plays a very huge roll in the in-house Prediction algorithm I built singlehandedly using the AI.
The algorithm also does not rely on match odds (bookmaker odds) to make predictions as the predictions are entirely precomputed on the backend using 4 stages factor plus a Bayesian model.
Therefore, the cron job I mentioned in my previous response wasn't just one, but rather a superset of crons that does exactly what you mentioned (fetching datas from an external source) as well as automating other backend security and match analysis integrity functions.

Auracletrader:
I know, it will sound very unbelievable but friends, I truly really did built a very beautiful, scalable and sophisticated AI Predictions web app that's very responsive without a single line of code.
It started as a curiosity to build and design a personal machine that can scan football matches everyday, analyze them, score them and then filter out the best by ranking with confidence scores that's based on so many metrics.

The tough begining:
I was very skeptical at first, like, hmm, will this AI thing build another AI powered thing?
Well, it sure did, only that it has to be refined and perfected with the use of precise prompting.

It should have taken me around 1 week to complete the MVP and ship it to production, but I wasn't always satisfied with what I have achieve within that kind of short time, so I refined and refined and refined, and 7 days turned two weeks, two weeks turned 1 month and a month turned 8 months.
But eventually, I got something beautiful only the best of the best software developers can achieve.

Look at it...isn't it beautiful?!
It's still in it's early stage, let's say beta.
In few weeks to come, I'll have added so many brilliant features to help all sports fans around the world understand the depth of the game better❤️

If you have any remarks, words of encouragement, criticism or questions regarding this, you are free to let me know.

And oh, since the web app has moved from an idea to a real thing, with around 890 daily users, then If anyone is interested in partnering or investing in it's marketing and growth, please reach out to heyzilobet@gmail.com

Impressive.... Which AI did you use?

Since U were using AI how come it took 8 months?

Wow!More grace sir.

This is impressive work. 890 daily users with no paid marketing is solid — most apps never get past the "my friends and family" phase.

The fact that you spent 8 months refining instead of shipping a rough MVP in one week says a lot. People underestimate how much iteration matters. AI can generate code fast, but getting the product right — the UX, the data pipeline, the confidence scores — that still takes human judgement and patience.

Curious about one thing: how do you handle the accuracy tracking? Like, do you have a system that compares your predictions against actual match results so users can see your hit rate over time? That kind of transparency is what builds trust in prediction tools.

Nice project overall. Keep building.

as69:
This is impressive work. 890 daily users with no paid marketing is solid — most apps never get past the "my friends and family" phase.

The fact that you spent 8 months refining instead of shipping a rough MVP in one week says a lot. People underestimate how much iteration matters. AI can generate code fast, but getting the product right — the UX, the data pipeline, the confidence scores — that still takes human judgement and patience.

Curious about one thing: how do you handle the accuracy tracking? Like, do you have a system that compares your predictions against actual match results so users can see your hit rate over time? That kind of transparency is what builds trust in prediction tools.

Nice project overall. Keep building.

HI, thanks for your input. It'll go a long way in motivating my spirit.
Yes, I've built an algorithm that use past matches, heat maps and player records to make impressive picks on individual matches. I connected the app to an external live scores and prematch data provider thaf feeds the algo with all the necessary data it needs to generate undeniably commendable tips and picks.