I saw a KOL on my TL talking about how he built an AI prediction market bot with an 80% win rate and I thought to myself why didn't @xomarket have this?
I reached out to him asking for help and directions but there was no response so I decided to build it myself and from scratch. Something to help new traders joining the community who needed a little assistance while placing trades🫂
The idea sounded simple, but the execution? It humbled me immediately.
-
Day 1 was pure chaos
I had no proper coding experience. So, I started from the very beginning.
1) Installing Node.js.
2) Setting up VS Code.
3) Learning what a terminal even was.
Every step had a problem:
• Node.js wasn't recognized after installation
• npm commands were getting cut off
•Git wasn't installed → another detour
•Gemini API hit daily quota limits so I switched to Groq
•The Groq model I switched to got deprecated
Every fix uncovered another problem behind it.
At multiple points I genuinely thought it would be where I stop but somehow I kept finding solutions one problem at a time
-
Day 2 was the turning point
After hours of debugging and tutorials, the bot finally responded inside Telegram, the bot finally responded.
I typed "/addmarket" in Telegram with a real XO Market question and got this back:
🔮 XO Market Forecast
📌 Delhi Capitals vs CSK - Total wickets exceed 9.5
📊 YES: 50% | NO: 50%
🤖 Most likely outcome: 9 or More Caught Wickets
Probability: 60% likely
Confidence: Medium
I just stared at my screen.
A bot I built from scratch without a proper coding background had just analyzed a live prediction market and generated a real AI forecast.
That moment genuinely changed something in my brain🤭
-
Day 3 was deployment
But there was a problem. Every time I closed my laptop, the bot went offline.
After a bit of research and that's when I discovered @Railway, a cloud hosting platform. After pushing my code to GitHub and deploying it there, the bot began running permanently in the cloud.
24/7. No laptop required.
-
What the Bot Does Today
✅ Analyzes any prediction market you feed it
✅ Generates AI forecasts with probability estimates
✅ Gives 3 key reasons for every prediction
✅ Shows live countdown timers on each market
✅ Sends automatic alerts every 6 hours
✅ Runs permanently in the cloud
✅ Accessible to anyone on Telegram
So what are the current limitations?
⚠️ The AI currently forecasts based on general knowledge, not live data
⚠️ XO Market doesn't have a public API yet so markets are added manually
⚠️ Forecast accuracy is still being improved
But here's what's coming Next.
• The next upgrade is significant, because the bot will read the actual source links provided for each market; news articles, stats pages and live data before generating its forecast.
• Real data going in means real accuracy coming out.
After that,
• Auto-posting forecasts to X/Twitter
• Full XO Market live API integration the moment it goes public
• Progressively smarter analysis over time.
This started as a random idea on a Tuesday, but now it's becoming a genuine tool for prediction market traders who want more clarity before committing their money.
The craziest part is that less than 2 weeks ago, I had never used a terminal properly in my life. Now I'm building something people can actually use and honestly, it's been one of the most rewarding things I've done.
Thank you for taking the time to read this 🩷
Deek has somehow acquired a plane & huge amounts of CP to give away.
No one approved this, but limited NFT boarding passes are now open to claim! There are four seats on this very questionable flight:
Economy: 1,000 CP
Business: 2,000 CP
First: 4,000 CP
Golden Ticket: 10,000 CP
Only one Golden Ticket exists.
Parlays are free to enter, powered by CP, and have already paid out $30,000+ in prizes over the last two months.
Get a pass to access Parlays. Link in comment 👇
• Cautious odds shift alerts
• Whale watching
• Resolution tracking
• Automa........
Currently working on this final lap before the forecast bot goes public. What an amazing journey so far 🩷
GM to those building daily.
As for me, i'll be spending the day working on telegram but first, let me give a run down of where we are now with the forecast bot
• Pulls live markets from @xomarket
• Finds the official resolution sources attached to each market
• Actually visits and reads those sources in real time
• THEN generates the forecast based on live information
But what happens when a website blocks the bot?
That's why i worked on a fallback system.
If:
✅ Source readable → AI uses live data → Better forecast
✅ Source blocked → AI uses live odds + general knowledge → Still solid
But what if the Bot fails silently?
That never happens, it always produces a forecast.
What's coming next:
• Markets sorted by urgency
• Search markets by topic or name
• Only HIGH confidence forecasts sent to subscribers on intervals
• Weekly top picks with outcome tracking
• Public accuracy score which will be fully automated with zero human interference
The goal is a self-improving prediction tool for @xomarket traders 🎯
So, let me know what you think can make this bot better. I'm also looking towards working with one person who can give consistent feedbacks on the product before the full release
@Judaixxspexxy@duality_daj Your first line shows you responded out of sentiment.
Read my comment again and understand that you're talking about a river when describing a forest
I stopped stressing myself 100% to write contents completely on my own. I did during citrea's so called testnet and even though projects building in the chain recognised my tweets, the chain and it's supposed Devs never did. I walked away with my wasted time, effort and data and ykw, no one even remembers today.
In conclusion, don't waste your time writing your articles 100% by yourself cause in the end, nobody in CT values creativity unless there's big numbers underneath. Them get why grok dey middle of your app. Don't lean in 100% but you xan use it to make your work easier. The time you might end up wasting isn't worth it, if your time don come to bang, na the one without much effort sef go trend
I stopped stressing myself 100% to write contents completely on my own. I did during citrea's so called testnet and even though projects building in the chain recognised my tweets, the chain and it's supposed Devs never did. I walked away with my wasted time, effort and data and ykw, no one even remembers today.
In conclusion, don't waste your time writing your articles 100% by yourself cause in the end, nobody in CT values creativity unless there's big numbers underneath. Them get why grok dey middle of your app. Don't lean in 100% but you xan use it to make your work easier. The time you might end up wasting isn't worth it, if your time don come to bang, na the one without much effort sef go trend
Oh, brother... I ain't even gonna lie, but staking money in a centralised prediction market is like throwing money in a casino with your eyes closed and hoping it lands on your pick.
Sorry about this but you should look towards decentralised prediction markets with locked sources and rules.
Why are schools still open in Nigeria?
Why are children still going to classes?
Over 40 children were abducted and NUT are still allowing pupils to come to school?
This Polymarket fiasco just proved how broken centralized resolution still is. Millions in bets are stuck because a key sale happened before June 1, but didn’t get publicized until after??!
And now Polymarket is doubling down and adding late info to backtrack a NO, despite the market clearly asking about that window.
And for a market that big, why should the resolution process be centralized, with their so-called judges whose wallets hold stakes? It just ruins the idea of a prediction market and becomes an obvious rigged casino.
Locked sources and Decentralised resolution is the future of prediction markets and anyone not doubling down on this concept would just plummet to 0
BREAKING: Polymarket has millions in bets stuck in dispute over one question: Did Strategy sell any bitcoin by May 31?
Strategy's own filing says yes, the sale ran from May 26 to 31. It just didn't go public until June 1, after the market closed. Now it all comes down to which date counts.
Resolution is the hardest problem in this whole category, and XO Market is built around it. The sources and the rules for reading them are locked when a market is created, so what counts is never up for debate after the close.
If you want to find out more, we outlined our approach in the article below.