From Bots to Web3: My First Project in Tokyo 🧩

In the summer of 2024, I landed in Tokyo for my first full-time project. I wasn’t handed a task—I was handed an opportunity. There was no roadmap, no handover. Just a short sentence:

"We need a Telegram Mini App. It has to go live in two weeks."

Most of the team wasn’t familiar with Telegram development, let alone how to turn it into a Web3-enabled engagement engine. But for me, this was familiar territory.

Why Me?

Before this, I had already built Telegram bots and managed channels with a few side projects. I understood the Telegram platform’s quirks—how it loads, what users expect, and how Mini Apps differ from traditional web apps.

And more importantly, this project wasn’t just any app. It combined two things I loved: Telegram and Web3.

The goal was ambitious but clear:

Build a daily check-in mini app to engage 4,000+ Web3 users — and launch it before November ends.

So I took the lead.

Why Telegram Mini Apps Are Trending in Tokyo

What surprised me was how relevant this was in Tokyo’s current tech scene. Companies here are in the middle of digital transformation (DX)—and many are looking for fast, low-cost, viral ways to reach users.

Telegram Mini Apps are perfect for this:

• 🧩 No app store delays. Users can launch your app instantly from a message.
• 🪄 Built-in notifications, perfect for daily habits like check-ins or rewards.
• 🖼️ Lightweight branding that still feels custom and mobile-native.

A few local startups in Shibuya were already experimenting with these—for NFT loyalty rewards, flash sales, even QR-based product check-ins. The viral “tap-to-earn” mechanic had proven sticky. So when we proposed our gamified check-in system, it felt right on time.

What I Actually Built

In just 2 weeks, I shipped an end-to-end Mini App that let users:

• ✅ Check in daily to earn points
• ✅ View leaderboards and personal stats
• ✅ Get real-time updates via Telegram push
• ✅ Sync with Web3 wallets to link rewards

I used ViteJS, TypeScript, FastAPI, and MySQL to keep things lightweight and fast. Deployed on Vercel, tested daily, and documented fully before handoff.

Behind the scenes, I wrote 30+ REST APIs and wired them into Telegram’s frontend context. The logic handled point tracking, user streaks, bot routing, and reward conditions—all under two weeks.

What Came Out of It

The impact wasn’t just technical. It was strategic.

• 🚀 We saw a +20% increase in daily active users (DAU) within a few weeks
• 👥 Over 4,000 Web3 users interacted with the app before the airdrop
• 🔁 Retention improved noticeably—before infrastructure costs even hit scale

And from a personal angle: it was the moment I realized I could own a product cycle end to end—from system design to delivery—and help shape how users feel and behave.

And That Was Just One Project...

Since then, I’ve gone on to

• 🧠 Auto-detect 15,000+ human settlements from satellite imagery using an AI-powered mapping pipeline — eliminating over 6 months of manual tagging
• ⛓️ Build a custom NFT marketplace with ERC20&ERC721 contracts — creating a new revenue stream via minting & secondary sales

But it all started here — with a Telegram app, a short deadline, and a team that trusted me to build something from scratch.

Curious about what’s next?

Follow me for more breakdowns, side projects, and adventures.

Or connect with me if you're building something crazy and need someone who knows how to turn ambiguity into launch day.