DEV Community: Rotji Gonsum

🚀 I Built an AI That Predicts When You’ll Fail Your Habits (MicroHabit AI)

Rotji Gonsum — Tue, 31 Mar 2026 19:39:46 +0000

By Rotji Gonsum
👋 Introduction
Most habit tracker apps help you track what you’ve already done.
But I kept asking myself:
What if an app could tell you when you’re about to fail… before it happens?
That idea led me to build MicroHabit AI — a habit tracking app that doesn’t just record your past, but predicts your future behavior.
👉Try it here

🧠 The Problem
Traditional habit trackers have one big limitation:
They show streaks
They log completions
They visualize progress
But they don’t actually help you avoid failure.
You only realize you’ve broken a habit after it happens.
That’s too late.

💡 The Idea
I wanted to build something different:
👉 An AI system that analyzes your habit patterns and predicts:

Which habit you’re likely to skip
When you’re most at risk

So instead of reacting to failure
👉 You prevent it.
🔥 The Core Concept
MicroHabit AI = Predictive Habit Tracking
“Know when you’re about to fail — before it happens.”

🛠️ Tech Stack
I built MicroHabit AI as a lightweight web app using:

React (PWA) – for fast, app-like experience
Firebase – for data handling and sync
AI APIs (Gemini) – for behavior predictions
Local Storage + Cloud Sync – hybrid data approach

🎨 UI/UX Journey
At first, my app looked… okay.
But it had a big problem:
👉 Users didn’t understand it immediately.
So I redesigned EVERYTHING:

Habits tab
AI Coach tab
Premium tab
Settings tab
Data tab
Landing page

⚠️ Biggest Lesson:
A beautiful UI means nothing if users don’t understand your product in 5 seconds.
📊 Early Analytics (First 30 Days)
After redesigning and sharing my app publicly:

57 active users (+128%)
56 new users (+115%)
6+ minutes average engagement (+129%)

🌍 Users came from:

USA
Nigeria
India
UK
Europe

🤯 What Surprised Me
People didn’t just open the app
👉 They stayed.
That told me something important:
The idea is interesting
The execution is improving
But clarity still matters the most
🧪 Building in Public (30 Days Challenge)
I decided to build this app in public for 30 days.
Every day I shared:

Progress updates
UI redesigns
Bugs and fixes

Lessons learned
💥 Example Problems I Faced:
AI predicting based on old data
Features that looked real but didn’t work
UI breaking immersion (default alert boxes)
Confusing onboarding
🧠 What I Learned:
Shipping fast is good
But clarity > features
Users don’t care about your effort
They care about understanding value instantly
💸 Monetization Plan
MicroHabit AI has 3 tiers:

🆓 Free (basic habits + limited AI)
💎 Premium ($9/month)
👑 VIP ($99/month / $999/year)

With features like:

Predictive analytics
AI coaching
Habit insights
Progress tracking

🚀 What’s Next
I’m preparing to launch on Product Hunt.
Before launch, I’m focusing on:
Improving onboarding
Simplifying messaging
Making the value crystal clear

🎯 Final Thought
Building this taught me something powerful:
The difference between a “cool project” and a real product is clarity.
🙌 Try It Yourself👉

💬 Let’s Connect
I’m sharing my journey publicly as I build this.
If you’re a developer, indie hacker, or just curious:
What do you think about predictive habit tracking?
Would you use something like this?

buildinpublic #react #saas #ai #webdev #productivity #indiehackers #startup

Day 19: I Built Breathing Glow Buttons (And They're Alive) 🌙

Rotji Gonsum — Mon, 23 Mar 2026 08:26:48 +0000

I'm on Day 19 of building MicroHabit AI in public. 30 days. Daily shipping. ProductHunt launch on March 29.

Today, I completely transformed the Habits tab from "functional" to "futuristic."

What is MicroHabit AI?

TL;DR: A habit tracker with AI coaching that predicts which habits you'll break and nudges you before you do.

The problem: Most habit trackers just track. They don't help you STAY consistent.

My solution:

📊 Track daily habits with streaks
🤖 AI analyzes your patterns and predicts failures
💬 Get personalized nudges BEFORE you break your streak
⏰ Smart reminders at optimal times
📅 Visual calendar showing your consistency
💀 Weekly "roast" feature (savage but supportive feedback)

Free tier: 5 AI predictions/day, current month calendar

Premium: Unlimited AI, full history, custom scheduling

Traction so far: 25 active users. 8 countries. 64% guest-to-signup conversion. $0 MRR (launching payments this week).

The Challenge

Yesterday, my UI looked... functional. It worked. But it didn't feel premium.

Users couldn't find features (they were hidden in a ⋮ menu). The calendar was plain HTML tables.
Browser alerts were ugly white boxes breaking the dark theme.

I needed to fix this before ProductHunt.

What I Shipped Today

✨ 1. Breathing Glow Buttons

The buttons literally pulse with life.

**Key decisions:
Multi-layer drop-shadows for depth
2.8s duration (feels natural, not rushed)
infinite alternate creates the breathing effect
Cyan + green for calendar, purple + violet for schedule
📅 2. Glowing Calendar History
Replaced HTML tables with a custom grid that shows:
✅ Completed days (green glow with pulsing animation)

❌ Missed days (subtle red highlight)
📆 Today (cyan ring)
.cal-day.done {
background: rgba(0,255,157,0.08);
border-color: rgba(0,255,157,0.5);
box-shadow: 0 0 14px rgba(0,255,157,0.2);
animation: glowGreen 2.5s ease-in-out infinite alternate;
}

@keyframes glowGreen {
from {
box-shadow: 0 0 10px rgba(0,255,157,0.2);
}
to {
box-shadow: 0 0 22px rgba(0,255,157,0.4);
}
}
The completed days literally breathe. It's mesmerizing.
💎 3. Dark Glassmorphism Modals
Goodbye ugly browser alerts. Hello custom modals:
.modal-content {
background: #0d0b1e !important;
backdrop-filter: blur(20px);
border: 1px solid rgba(0,229,255,0.3);
border-radius: 18px;
}
Now when you delete a habit or set a reminder, you get:
Dark themed modal (not white!)
Yes/No buttons (not just OK)
Color-coded by action (red for delete, cyan for confirm)
Smooth slide-up animation
🎯 4. Visible Action Buttons
Hidden ⋮ menu → Visible buttons with icons:
[⏰ Remind] [🎯 Goal] [📅 History] [📝 Note] [🗑️ Delete]

Each color-coded:
Purple for reminders
Cyan for goals
Green for history
Blue for notes
Red for delete
Mobile-friendly. No hover required. Discoverable.
Technical Stack
100% Vanilla:
No frameworks
Pure CSS animations (GPU accelerated)
Custom SVG icons
Responsive design
Why vanilla?
Fast (no bundle size)
Simple (no build step)
Educational (anyone can read the code)
Portable (works anywhere)
Performance
All animations run at 60fps because:
CSS animations (not JavaScript)
GPU accelerated (transform, filter)
No layout thrashing
Optimized drop-shadows
Lighthouse score: Still 95+ 🚀
What I Learned

Multi-layer drop-shadows create depth One shadow = flat. Four shadows = dimensional.
Breathing animations need the right timing Too fast (1s) = anxiety Too slow (5s) = boring Sweet spot: 2.5-3s
Color psychology matters Cyan/green = growth, success (calendar) Purple/violet = wisdom, planning (schedule)
!important isn't always evil When modals inherit conflicting styles, !important saves you.
Users need Yes/No, not just OK Browser confirms have no cancel. Custom modals do. Before → After Before: Hidden features (⋮ menu) Static design Ugly white alerts Plain HTML tables After: Visible action buttons Breathing animations Dark glassmorphism modals Glowing calendar grid What's Next Tomorrow or today (Day 20 or Day 19): AI Coach tab redesign March 29: ProductHunt launch 🚀 Live now: Try MicroHabit AI Questions for You What's your favorite CSS animation technique? Do breathing/pulsing effects annoy you or draw you in? Should I open-source the UI components? Building in public. Shipping daily. 10 days until launch. Follow along: Twitter: MicroHabitAI Linkedin: Fancy Treaz

See before and after

buildinpublic #30daychallenge #ProductHunt

Day 19: I Built Breathing Glow Buttons (And They're Alive) 🌙

Rotji Gonsum — Mon, 23 Mar 2026 08:26:48 +0000

I'm on Day 19 of building MicroHabit AI in public. 30 days. Daily shipping. ProductHunt launch on March 29.

Today, I completely transformed the Habits tab from "functional" to "futuristic."

What is MicroHabit AI?

TL;DR: A habit tracker with AI coaching that predicts which habits you'll break and nudges you before you do.

The problem: Most habit trackers just track. They don't help you STAY consistent.

My solution:

📊 Track daily habits with streaks
🤖 AI analyzes your patterns and predicts failures
💬 Get personalized nudges BEFORE you break your streak
⏰ Smart reminders at optimal times
📅 Visual calendar showing your consistency
💀 Weekly "roast" feature (savage but supportive feedback)

Free tier: 5 AI predictions/day, current month calendar

Premium: Unlimited AI, full history, custom scheduling

Traction so far: 25 active users. 8 countries. 64% guest-to-signup conversion. $0 MRR (launching payments this week).

The Challenge

Yesterday, my UI looked... functional. It worked. But it didn't feel premium.

Users couldn't find features (they were hidden in a ⋮ menu). The calendar was plain HTML tables.
Browser alerts were ugly white boxes breaking the dark theme.

I needed to fix this before ProductHunt.

What I Shipped Today

✨ 1. Breathing Glow Buttons

The buttons literally pulse with life.

Multi-layer drop-shadows create depth One shadow = flat. Four shadows = dimensional.
Breathing animations need the right timing Too fast (1s) = anxiety Too slow (5s) = boring Sweet spot: 2.5-3s
Color psychology matters Cyan/green = growth, success (calendar) Purple/violet = wisdom, planning (schedule)
!important isn't always evil When modals inherit conflicting styles, !important saves you.
Users need Yes/No, not just OK Browser confirms have no cancel. Custom modals do. Before → After Before: Hidden features (⋮ menu) Static design Ugly white alerts Plain HTML tables After: Visible action buttons Breathing animations Dark glassmorphism modals Glowing calendar grid What's Next Tomorrow or today (Day 20 or Day 19): AI Coach tab redesign March 29: ProductHunt launch 🚀 Live now: Try MicroHabit AI Questions for You What's your favorite CSS animation technique? Do breathing/pulsing effects annoy you or draw you in? Should I open-source the UI components? Building in public. Shipping daily. 10 days until launch. Follow along: Twitter: MicroHabitAI Linkedin: Fancy Treaz

See before and after

buildinpublic #30daychallenge #ProductHunt

Day 17: I Built Habit Reminders That Work Offline (The Debugging Hell Was Worth It)

Rotji Gonsum — Sat, 21 Mar 2026 23:02:08 +0000

Day 17/30: Service Worker Debugging Hell & Victory 🔥

TL;DR

Built offline reminders for MicroHabit AI
Spent 6 hours debugging Service Worker
Fixed repository path issues
Unlocked calendar for free users
25 users, 8 countries, high engagement

The Problem

Users wanted reminders. Easy, right?

Wrong.

Browser reminders only work when app is open.
Close the app = no notifications.

I needed Service Workers.

The Journey

Hour 1: "This should be easy"

Wrote service-worker.js.
Deployed.
Didn't work.

Hour 2-3: 404 Errors

Service Worker not found.
Path issues.
Repository name problems.

Hour 4-5: "Illegal Constructor"

PWAs can't use new Notification()
Must use registration.showNotification()
Who knew?!

Hour 6: Victory

Changed notification method.
Fixed paths.
IT WORKS! 🎉

What I Learned

1. Service Workers are hard
But worth it for offline functionality.

2. PWA != Website
Different APIs, different rules.

3. User education matters
Added tips: "Keep app in background"
Not perfect, but transparent.

4. Debug panels save lives
Built on-screen debugging.
No more guessing on mobile!

The Stats

Current MicroHabit AI metrics:

25 active users
8 countries
30 events/user
60% weekly growth
$0 MRR (haven't monetized yet!)

What's Next

Tomorrow (Day 18):

Time suggestions per habit
Grow user base
Start Preparation for ProductHunt launch

Try It

Link: MicroHabit AI

Free tier: 3 habits, AI predictions
Premium: $9/month, unlimited everything

Code

Full build documented on GitHub
Building in public from Nigeria 🇳🇬

Questions? Drop them below! 👇

Follow the journey:

Twitter: [MicroHabitAI]
LinkedIn: [Fancy Treaz]

buildinpublic #javascript #pwa #saas

Taking 3 MVP Projects Built My SaaS in 16 Days on a Phone (Proof Inside)

Rotji Gonsum — Thu, 19 Mar 2026 23:01:21 +0000

**Update: **2 spots left! (First client secured)

After spending 16 days building MicroHabit AI on my phone and documenting the journey here on Dev.to, I'm opening up to take *3 MVP projects.
*

What I Just Built

MicroHabit AI - An AI-powered habit tracker that predicts which habits you'll fail before you actually fail them.

Tech Stack:

Frontend: HTML/CSS/JavaScript
Backend: Firebase (Authentication, Firestore, Cloud Functions)
AI: Gemini API integration
PWA: Installable on any device
Hosting: GitHub Pages

Timeline: 16 days from idea to live app

Built entirely on my phone in Jos, Nigeria.

Try it live: https://microhabitai.github.io/MicroHabit-AI/

Read the journey: Check my previous posts - I documented every day, every bug, every decision.

What I'm Offering

I'll build your MVP in 30-45 days for ₦150k-₦300k ($200-$400 USD).

What You Get:

✅ Fully functional web app

Responsive design (works on all devices)
User authentication
Database integration (Firebase)
PWA capabilities (installable like native app)

✅ Fast delivery

30-45 days from agreement to launch
Daily progress updates
Weekly demos

✅ Clean, maintainable code

Documented
Scalable architecture
Easy to hand off to your team later

✅ Optional add-ons:

AI integration (+₦50k)
Payment integration (+₦50k)
Admin dashboard (+₦30k)
Mobile app version (+₦80k)

Perfect For:

🎯 Startup founders who need MVPs to validate ideas

🎯 Non-technical entrepreneurs with great ideas but no tech team

🎯 Businesses that need simple web apps fast

🎯 Side projects that need to launch quickly

🎯 Anyone who wants proof their idea can work before investing heavily

Why Work With Me?

1. Proven Speed
I shipped MicroHabit AI - a full SaaS with AI integration - in 16 days. Your MVP will be simpler, so I can deliver even faster.

2. Affordable
I'm building my portfolio while building MicroHabit AI. You get quality work at fair rates.

3. Transparent Process
I document everything. You'll see progress daily. Check my Dev.to posts to see how I work.

4. Real Proof
Most developers show you GitHub repos. I'm showing you a live, working app that people are actually using.

5. Located in Nigeria
Easy timezone collaboration. Same currency. Understanding of local market needs.

My Tech Stack

Frontend:

HTML5, CSS3, JavaScript (ES6+)
Responsive design
PWA capabilities

Backend:

Firebase (Firestore, Authentication, Cloud Functions)
RESTful APIs
Serverless architecture

Integrations:

Payment gateways (Paystack, Flutterwave)
AI APIs (Gemini, OpenAI)
Email services
Analytics

Deployment:

GitHub Pages
Netlify
Vercel
Firebase Hosting

Pricing Packages

🚀 Quick MVP - ₦150,000 (30 days)

Simple web app (5-7 pages)
User authentication
Basic database
Responsive design
PWA setup

⚡ Full MVP - ₦250,000 (45 days)

Most Popular

Everything in Quick MVP
Advanced features
Admin dashboard
Email notifications
Analytics integration
2 rounds of revisions

🔥 Premium MVP - ₦350,000 (60 days)

Everything in Full MVP
AI integration
Payment integration
Mobile app version (PWA)
Launch support
3 rounds of revisions

Only Taking 3 Projects

I'm limiting to 3 projects so I can:

Give each project proper attention
Maintain quality
Continue building MicroHabit AI
Document the process

First come, first served.

How to Apply

Send me a DM here on Dev.to or email: fancytreaz@gmail.com

Include:

Your name and business/project name
Brief description of your idea (2-3 sentences)
Key features you need (bullet points)
Timeline expectations
Budget range
Why this project matters to you

I'll respond within 24 hours with:

Initial thoughts on feasibility
Rough timeline estimate
Whether I'm a good fit for your project

What Happens After You Apply?

Step 1: You send inquiry → I respond within 24 hours

Step 2: 15-minute discovery call (free) → Understand your needs

Step 3: I send detailed proposal → Timeline, features, cost

Step 4: You approve → 50% deposit, work begins

Step 5: Weekly progress updates → You see your app being built

Step 6: Final delivery → 50% final payment, app is yours

Step 7: 30 days support → Bug fixes, minor adjustments

FAQs

Q: Can you build [specific thing]?
A: If it's a web app, probably yes. DM me and let's discuss.

Q: Do you work with international clients?
A: Yes! Pricing in USD: $200-$400 depending on complexity.

Q: What if I need changes after delivery?
A: First 30 days: Free bug fixes and minor tweaks. Major changes quoted separately.

Q: Can I see more of your work?
A: MicroHabit AI is my latest. I've built internal tools for previous employers (NDA prevents sharing). This is my first public portfolio piece.

Q: Why so affordable?
A: I'm building my portfolio. As I get more clients and testimonials, rates will increase. Lock in current rates now.

Q: Payment terms?
A: 50% upfront, 50% on delivery. Nigerian clients: Bank transfer. International: Wise/PayPal.

About Me

I'm Rotji Gonsum (Treaz), a developer in Jos, Nigeria.

I built MicroHabit AI in 16 days on my phone while documenting the entire journey here on Dev.to.

Why did I build on my phone? Because that's what I had. And I proved it's possible.

Now I'm helping other founders and entrepreneurs build their MVPs fast and affordably.

Currently: Day 16 of 30-day build-in-public challenge
Next: Product Hunt launch (Day 30)
Location: Jos, Nigeria 🇳🇬

Connect:

Dev.to: You're already here!
LinkedIn: [Fancy Treaz]
Twitter/X: @MicroHabitAI
Live Project: https://microhabitai.github.io/MicroHabit-AI/

Let's Build Something Together

I have 3 slots open this month.

If you've been sitting on an idea, waiting for the right developer, waiting for the right price, waiting for proof it can be done...

This is it.

DM me. Let's talk.

First 3 serious inquiries get priority.

Tags: #mvp #freelance #webdev #firebase #javascript #nigeria #startup #buildinpublic

P.S. - If you're reading this and thinking "I wish I could build like this" - you can. I documented everything. Start with Day 1 of my MicroHabit AI journey. All the code, all the decisions, all the struggles. It's all here.

But if you need it built NOW, that's what I'm here for. 🚀

Day 15/30: Building MicroHabit AI - How I Broke My Notification System 3 Times in 6 Hours

Rotji Gonsum — Thu, 19 Mar 2026 10:12:56 +0000

Rotji Gonsum's journey building MicroHabit AI, an AI-powered habit tracker, solo on a phone in Nigeria. Day 15 of 30-day build-in-public challenge.

Day 15/30: Building MicroHabit AI - How I Broke My Notification System 3 Times in 6 Hours

*By Rotji Gonsum (Treaz), Solo Founder of MicroHabit AI
*
I'm Rotji Gonsum, a solo developer building MicroHabit AI an AI-powered habit tracker that predicts which habits you'll fail before it happens. I'm building this entirely on my phone in Nigeria as part of a 30-day build-in-public challenge.

This is Day 15. Here's what happened when I spent 6 hours (11pm to 5am) building a notification system that kept breaking.

About MicroHabit AI

Before I dive into the Day 15 chaos, let me tell you what MicroHabit AI actually does.

MicroHabit AI is not another habit tracker. It's an AI-powered predictive coaching system that analyzes YOUR specific patterns and tells you which habits you're about to fail - before you actually fail them.

Unlike apps like Habitica or Streaks that just track what you did, MicroHabit AI uses AI to predict what you WON'T do, then coaches you through it.

I'm Rotji Gonsum (also known as Treaz online), and I'm building MicroHabit AI solo on my phone in Jos, Nigeria. Why my phone? Because that's what I have. No fancy MacBook. Just an Infinix phone, borrowed data, and determination.

You can try MicroHabit AI here: https://microhabitai.github.io/MicroHabit-AI/

*The Day 15 Mission: Build Notification System
*
Based on feedback from Angela (a UX expert who replied to my LinkedIn InMail), I learned that the biggest problem with habit trackers isn't the features, it's that people forget to open them.

So Day 15's goal was simple: Build a notification system for MicroHabit AI that reminds users to check their habits.

Not just generic "check your habits!" reminders. But personality-driven notifications based on how users WANT to be motivated.

I designed 4 personality styles:

💪The Motivator Encouraging and positive -** 📊 The Realist** - Data-driven and analytical
💀 The Savage - Brutally honest but caring
🎯 The Challenger - Competitive and bold

Smart idea, right?

*The 6-Hour Debugging Marathon
*
Here's the timeline of my all nighter building the MicroHabit AI notification system:

11:00 PM - Started Coding

Fresh night data plan. Full of optimism. Started building the personality-driven banner system for MicroHabit AI.

11:30 PM - First Version Complete

Built 4 different notification personalities. Each with 3 variations (morning, afternoon, evening). That's 12 unique messages for MicroHabit AI users.

Code looks clean. Testing locally works perfectly.

1:00 AM - Integrated Into Onboarding

I thought: "Let users choose their personality style during onboarding! They'll love customizing MicroHabit AI to match their preferences!"

Added a personality picker to the onboarding modal. Users had to choose before using the app.

Seemed logical.

1:30 AM - First Bug

Opened MicroHabit AI on my phone to test.

The personality picker appeared. But it wouldn't go away.

Even worse - it blocked the entire landing page. New visitors couldn't even SEE what MicroHabit AI does. They just saw this popup asking them to choose a motivation style for an app they haven't even tried yet.

Terrible UX.

2:00 AM - Fixed Auth Logic

Added authentication checks. The personality picker should only show for logged-in users, not random visitors to MicroHabit AI's landing page.

Tested. Worked!

Or so I thought.

2:30 AM - Second Bug

Logged into MicroHabit AI as an existing user.

The personality picker appeared. Again.

Clicked through, selected "Savage" style. Modal closed.

Refreshed the page.

It appeared AGAIN.

The notification system in MicroHabit AI was showing the personality picker EVERY SINGLE TIME users opened the app. Even after they'd already chosen.

3:00 AM - Fixed localStorage Logic

Added localStorage tracking. Mark when user has chosen their personality style. Don't show picker again.

Should work now, right?

4:00 AM - Testing on Web Browser

Opened MicroHabit AI in Chrome.

Still broken.

The personality picker showed every time. Even with the localStorage fix.

4:30 AM - User Report

Got a message: "The notification thing shows up every time I open the app. Even on the installed version."

Now it's affecting both web AND PWA versions of MicroHabit AI.

Made it worse.

5:00 AM - The Nuclear Option

Data expires at 6 AM. I'm tired. The feature is broken in three different ways.

I opened the code for MicroHabit AI and did something drastic:

I deleted 60 lines of code.

Removed the personality picker from onboarding entirely.

But first, I checked how successful apps handle this.

*What Duolingo, Headspace, and Habitica Taught Me
*
At 5am, exhausted and frustrated with MicroHabit AI's broken notification system, I checked how the pros do it:

Duolingo: No customization in onboarding. Uses sensible defaults. Lets users customize later in Settings.
Headspace: Simple onboarding. Default notification style. Settings for preferences.
Habitica: Just works out of the box. Customization is optional, not forced.
Notion: Sensible defaults first. Advanced settings for power users.

Pattern spotted: Nobody puts preference choices in onboarding.

They all use smart defaults and let users customize LATER if they want.

The Fix for MicroHabit AI

I simplified the MicroHabit AI notification system:

Removed personality picker from onboarding
Auto-set default to "Motivator" (positive, works for everyone)
Moved customization to Settings tab
Kept all 4 personality styles available

Result:

✅ Clean onboarding
✅ No bugs
✅ Works immediately
✅ Can still customize in MicroHabit AI Settings

Tested at 5:30 AM. Worked perfectly.

The notification system for MicroHabit AI was finally live.

*Testing the MicroHabit AI Notification System
*
This morning (10 AM), I opened the installed PWA version of MicroHabit AI.

A banner slid in from the top:

"🎯 Challenge! Complete at least 3 habits today. Prove you can do it!"

The Challenger personality style. Because I changed it in Settings to test.

It worked. The MicroHabit AI notification system was alive.

What I Learned Building MicroHabit AI (Day 15)
**
**Lesson 1: Sometimes the best code is deleted code.

60 lines removed from MicroHabit AI made it better than 60 lines added.

Lesson 2: Copy what works.

Duolingo, Headspace, and Habitica spent millions figuring out onboarding. I can learn from them instead of reinventing the wheel for MicroHabit AI.

**Lesson 3: **Defaults matter more than options.

Most MicroHabit AI users will never change the default Motivator style. And that's okay. It works for 90% of people.

Lesson 4: Don't force users to configure things.

The best features in MicroHabit AI are the ones that just work.

*Current Status of MicroHabit AI
*
Day 15 of 30 complete.

Users: Still 16 (stuck for a week)
Features shipped: Notification system with 4 personality styles
Lines of code deleted: 60
Hours of sleep: 2
Days until Product Hunt launch: 16

MicroHabit AI now has:

✅ AI-powered habit predictions
✅ Weekly roasts (savage but supportive feedback)
✅ Streak tracking with fire emojis
✅ Personality-driven notifications (NEW!)
✅ One-tap habit logging

Still building solo. Still on my phone. Still in Nigeria.

Try MicroHabit AI

Want to see the notification system in action?

Try MicroHabit AI: https://microhabitai.github.io/MicroHabit-AI/

Free version: 3 habits, 10 AI predictions per day
Premium: Unlimited everything

Built by Rotji Gonsum (Treaz) in Jos, Nigeria.

Follow my build-in-public journey:

Dev.to: @Rotji Gonsum
LinkedIn: Fancy Treaz
Twitter/X: @MicroHabitAI

The Question

*What's a feature you spent hours building only to realize simpler was better?
*

Have you ever deleted code and made your product better?

Let me know in the comments. I'm learning as I build MicroHabit AI, and I'd love to hear your stories.

This is Day 15 of my 30-day challenge building MicroHabit AI in public. Follow along to see a solo founder build an AI-powered habit tracker from scratch on a phone in Nigeria.

About the Author:
Rotji Gonsum (Treaz) is a solo developer building MicroHabit AI - an AI-powered habit tracker that predicts habit failures before they happen. Building entirely on a phone in Jos, Nigeria. Day 15 of 30-day build-in-public challenge leading to Product Hunt launch.

Try MicroHabit AI:
MicroHabit AI

Day 14/30: Got 0 Impressions, 0 Signups, and 3 Messages Worth More Than Gold POST:

Rotji Gonsum — Wed, 18 Mar 2026 23:30:51 +0000

Day 14/30: Got 0 Impressions, 0 Signups, and 3 Messages Worth More Than Gold

I'm going to be brutally honest about Day 14.

The Numbers (They Suck)

What I did today:

Posted on Twitter
Posted on LinkedIn
Sent 10 cold InMails
Kept building MicroHabit AI

What I got:

Twitter impressions: 0
LinkedIn impressions: 24
New signups: 0
Users: Still 16 (stuck here for a week)

After 14 days of daily shipping, those are rough numbers.

Most people would look at this and think: "This isn't working. Time to quit."

But Here's What Actually Happened

Three people replied to my LinkedIn InMails.

Not just polite "thanks but no thanks" replies.

Real, thoughtful responses from people who actually GET productivity and user behavior.

Let me share what they told me.

Reply 1: The Personality Problem
*From Favour (Productivity Coach):
*

"The biggest gap in habit trackers is personalization based on personality type. Most trackers assume everyone's brain works the same way. They push daily streaks (which work for some, but paralyze others). They use rigid tracking (which helps structure-oriented people, but suffocates flexibility-oriented people)."

Mind. Blown. 🤯

I've been building ONE approach for everyone.

But some people need:

Flexible weekly goals (not rigid daily)
Social accountability (not solo tracking)
Gentle suggestions (not harsh streaks)

She just explained why my conversion rate is 0%.

Reply 2: The Retention Problem

From Angela (UX Expert):

"Consistent time management is most people's problem. Getting people to use apps - a plan or something built in to remind them to use it. A text reminder or better, something to show up on the screen."

She looked at MicroHabit AI and found the elephant in the room:

I'm building a habit tracker that requires a habit of using the habit tracker.
😅

The irony is brutal.

People download it. Then forget it exists.

MicroHabit AI doesn't have notification reminders yet. I've been so focused on AI predictions and roast features that I forgot the obvious:

If people don't open it, none of my features matter.

Angela just saved me from building a beautiful product nobody remembers to use.

Reply 3: The Balance

From Jaiye (Parent):

"Habit trackers are very individual things. Some people like to keep them simple (a grid to tick), while others need something interactive (like Duolingo does)."

Simple vs interactive.

I've been trying to be both.

Maybe that's why it's confusing?

Maybe I need a "Simple Mode" toggle for people who just want checkboxes?

What I Learned

Here's the thing about those 0 impressions and 0 signups:

They tell me my DISTRIBUTION is broken.

But those 3 replies tell me my PRODUCT has real problems to solve:

Not personalized enough (Favour's insight)
No retention system (Angela's insight)
Trying to please everyone (Jaiye's insight)

I'd rather have 3 brutally honest replies than 1000 users who quietly leave.

Because now I know what to build.

What I'm Doing Next

Day 15-17:
Building notification system based on Angela's feedback.

Not generic "check your habits!" reminders that get ignored.

But:

"Your 7-day streak is at risk 🔥" (anxiety-driven)
"Your weekly roast is ready 💀" (curiosity-driven)
"AI predicts you'll skip today... prove it wrong" (challenge-driven)

Day 18-20:
Adding personality-based customization based on Favour's feedback.

Day 21-30:
Polish for Product Hunt launch.

The Metrics vs The Learning

Today's metrics: Terrible

Today's learning: Invaluable

I'm building MicroHabit AI - a habit tracker with AI that predicts which habits you're about to fail.

16 users. Built entirely on my phone in Nigeria.

16 days until Product Hunt.

The numbers will come. But first, I need to build something people actually want to keep using.

**Try it and tell me what's broken

What's the gap between your metrics and your learnings?

Drop a comment - I read every one.

I Built a SaaS App on a Broken Phone That Got Stolen - Here's What I Learned

Rotji Gonsum — Sat, 14 Mar 2026 17:20:42 +0000

By Rotji Gonsum (Fancytreaz)

Location: Jos, Nigeria 🇳🇬

App: MicroHabit AI - AI-Powered Habit Tracker

The Beginning

January 2026. I had an idea for a habit tracker that could predict when you'd skip your habits before it happened. Not another reminder app. Real AI predictions based on your patterns.

I had:

A broken Android phone (crashed every 10 minutes)
Zero budget
No MacBook
No "proper" development environment
Just QuickEdit (a text editor for Android)

Most people would have waited for "better circumstances."

I started building anyway.

Building on a Broken Phone

Every coding session was a race against time.

Write code for 8 minutes. Phone crashes. Restart. Push to GitHub. Repeat.

People asked: "Why don't you just get a laptop?"

The answer was simple: I couldn't afford one.

So I built with what I had.

Tech stack:

Vanilla JavaScript (no frameworks - couldn't install npm on my phone)
Firebase Firestore (free tier)
Firebase Authentication (free tier)
Google Gemini API (free tier)
GitHub Pages (free hosting)
Total cost: $0

The First Users

February 3, 2026 - Launch day.

I posted on LinkedIn. Shared on Twitter. Told everyone I knew.

Day 1:3 signups

Day 7:8 users

Day 14:14 users, 5 countries

People from Bulgaria, Ireland, Netherlands, USA, and Nigeria were using something I built on a phone that crashed every 10 minutes.

That felt incredible.

Then Everything Changed

February 17, 2026.

My phone got stolen.

The broken phone I'd spent a month coding on. Gone.

I borrowed my neighbor's phone to post updates. "Phone stolen. App still running."

The responses shocked me:

"Shipping really is the moat"
"This is what building in public looks like"
"Your dedication is inspiring"

The lesson:They took my hardware. They couldn't take what was deployed.

The Comeback

March 3, 2026.

My dad bought me a new phone - Infinix Smart 10 HD.

First thing I did: Clone my GitHub repo.

The app was exactly as I left it. 15 users. 5 countries. Zero downtime.

I'd gained 2 new users while my phone was stolen.

The internet doesn't care what device you build on.

The Reality Check

One month in. 15 users. Zero paid.

Everyone was using the free tier (3 habits). Nobody upgrading to premium ($9/month).

Honest truth:Most of my users are broke. I know some of them personally. They can't afford $9/month.

This hurt. But it taught me something important:

Building for people like you means understanding their constraints.

I'm not building for Silicon Valley developers with $200k salaries. I'm building for people in Jos, Nigeria. People coding on broken phones. People with $0 budgets.

That's my market. That's who I understand.

What I've Learned

Lesson 1: Perfect Conditions Are a Myth

Waiting for a MacBook? A standing desk? Dual monitors? You'll wait forever.

Start with what you have. The gap between "I wish I had" and "I'm building with" is where real builders live.

Lesson 2: Deployment > Development Environment

My code wasn't written in VS Code with Copilot. It was written in QuickEdit on a cracked screen.

But it shipped. It's live. People use it.

Nobody asks what IDE you used. They ask if it works.

Lesson 3: Free Users ≠ Failed Product

15 free users, 0 paid sounds like failure.

But those 15 people chose my app over Habitica, Streaks, and dozens of established competitors.

They're not paying because they CAN'T, not because the product isn't valuable.

Different problem. Different solution needed.

Lesson 4: Building in Public Creates Opportunities

Posting my journey led to:

Real connections with other builders
Advice from experienced founders
Features in tech blogs
Discord communities welcoming me
One incredible comment: "I coded on my dad's phone for 4.5 years - you have potential"

I didn't get users from building in public.

I got something better: a community.

Lesson 5: Your Story IS Your Marketing

I don't have a marketing budget.

But I have a story:

Built on broken phone
Phone got stolen
App kept running
Built from Nigeria
Still here

People remember stories. Nobody remembers "just another habit tracker."

Lesson 6: Constraints Breed Creativity

No npm? Write vanilla JavaScript.

No database money? Use Firebase free tier.

No hosting budget? GitHub Pages.

No API budget? Find free tiers.

I learned more building with constraints than I would have with unlimited resources.

Lesson 7: Most People Quit at Day 3

I'm on Day 4 of a 30-day posting challenge.

Day 1: Excited, full of energy

Day 2: Still motivated

Day 3: Reality sets in, results are slow

Day 4: Most people quit here

The ones who succeed? They show up on Day 4.

The Numbers (Honest Reality)

After 1 month:

15 registered users
5 countries reached
0 paid users
$0 revenue
$0 spent on marketing
~200 total video views
100% uptime

Are these impressive numbers? No.

Are they honest numbers? Yes.

What's Next

I'm not giving up on my 15 users.

I'm not chasing vanity metrics.

I'm doing two things:

Posting daily for 30 days:

Building consistency. Building content. Building in public.
Launching on Product Hunt (March 29): One big swing. One day of massive traffic. One chance to 10x my users.

Will it work? I don't know.

But I'm showing up.

The Real Success

Success isn't 1,000 users or $10k MRR yet.

Success is:

Building something people use (15 people trust my code)
Surviving setbacks (phone stolen, kept going)
Learning constantly (every bug is a lesson)
Showing up daily (Day 4 of 30)
Staying honest (no fake metrics, no exaggerations)

*To Anyone Building with Limited Resources
*
You don't need:

A MacBook
A standing desk
The latest iPhone
A coding bootcamp certificate
Silicon Valley connections
A marketing budget

You need:

An idea
Internet access
The willingness to start
The discipline to continue

I'm not special. I just started.

With a broken phone that got stolen.

And I'm still here.

Current Status

App: Live at click here
Users: 15

Revenue: $0

Streak: 4 days of daily posting

Next milestone: Product Hunt launch (March 29)

Current mood: Determined

Connect With Me

Twitter/X: @MicroHabitAI

LinkedIn: Fancy Treaz
Location: Jos, Nigeria 🇳🇬

Status: Building in public, one day at a time

If you're building something with limited resources, let's connect. We're in this together.

Day 4 of 30. Still here. Still building.

This post was written on an Infinix Smart 10 HD, the phone my dad bought me after my previous one was stolen. The journey continues.

I Built a SaaS App on a Broken Phone with Zero Budget - Here's What Happened

Rotji Gonsum — Sat, 14 Feb 2026 10:59:49 +0000

A story about constraints, stubbornness, and shipping anyway

Six months ago, if you told me I would build and launch a real SaaS product, I would have laughed.

Not because I didn't have ideas. I had plenty of those.

But because I had none of the things everyone says you need to build software:

No laptop
No money
No coding bootcamp
No team
No investors
No proper development environment

What I did have was a broken Android phone, a stubborn refusal to quit, and Internet connection.

This is the story of how I built MicroHabit AI from Jos, Nigeria and what I learned along the way.

The Idea

I wanted to build something that would help people stay consistent with their habits.

Not another app that just sends you notifications you ignore. Something smarter. Something that could actually predict when you were about to fail - and help you before it happened.

The idea was simple: use AI to analyze your personal habit patterns and give you a nudge at exactly the right moment.

I had no idea how hard it would be to build this on a cracked screen that crashed every 10 minutes.

The Setup (Or Lack of It)

Let me paint the picture of my development environment:

*My "computer": A broken Android phone with a cracked screen
*My code editor: QuickEdit - a free mobile app
*My backend: Firebase - Google's free tier
*My hosting: GitHub Pages - completely free
*My payments: Paystack
*My AI: Google Gemini API - free tier
*My tutorials: YouTube at 2AM when the house was quiet
*My budget: $0

Every developer tutorial I found assumed I had a laptop. Every Stack Overflow answer assumed I had a terminal. Every YouTube video showed someone with a proper setup.

I had none of that.

So I had to figure everything out differently. When you can't copy someone else's setup, you have to actually understand what you're building. That turned out to be the most valuable constraint I had.

The Building Process

I won't pretend it was smooth.

Here's what a typical building session looked like:

Open QuickEdit on broken phone
Write code for 30 minutes
Phone overheats and crashes
Restart phone
Lose some unsaved progress
Write it again (better this time)
Test in mobile browser
Find a bug
Fix the bug
Create two new bugs
Fix those
Repeat

Debugging on a 5-inch cracked screen is an experience I wouldn't wish on anyone. But it also meant every line of code I wrote, I understood completely. There was no room for copy-pasting things I didn't understand.

The phone crashing constantly actually made me a better developer. I learned to save frequently. I learned to write smaller, cleaner functions. I learned to test as I go.

Your constraints shape your craft.

What I Built

MicroHabit AI is an AI-powered habit tracker with a specific focus: predicting failure before it happens.

Here's how it works:

You add your habits and set weekly goals
You check off completions every day
The AI (powered by Google Gemini) analyzes your patterns
It predicts which habits you're likely to skip tomorrow
It gives you a personalized nudge to prevent it

Features that made it into the final product:

*AI Predictive Coaching - Real predictions based on your personal data
*Streak Tracking - Daily and weekly streaks with progress bars
*Calendar History - Full visual history of every habit
*Custom Goals - Set 1-7 days per week per habit
*Cloud Sync - Works across all your devices
*Any Timezone - No broken streaks when traveling

Pricing:

Free plan: Up to 3 habits
Premium: $9/month for unlimited everything
The Tech Stack (All Free)

For the developers reading this, here's exactly what I used:

*Frontend: Pure HTML, CSS, and JavaScript (no frameworks)
*Database: Firebase Firestore
*Authentication: Firebase Auth
*Hosting: GitHub Pages
*Payments: Paystack
*AI: Google Gemini API
*Code Editor: QuickEdit (Android app)
*Version Control: GitHub mobile app

Total monthly cost to run: $0 (within free tiers)

The fact that all of this is free in 2026 is genuinely remarkable. The barriers to building software have never been lower. You do not need money to start.

Launch Day

I launched on February 3rd, 2026.

No big announcement. No press coverage. No Product Hunt launch. Just pushed the code to GitHub Pages and shared the link on social media.

The first signup came within hours.

I cannot describe what that felt like. Someone, somewhere in the world, found my app and decided to create an account. A real person. On something I built on a broken phone.

The Numbers (Honest and Unfiltered)

I believe in radical transparency so here are the real numbers 11 days after launch:

*Total users: 14
*Paid users: 0
*Payout received: Yes (from a test payment I made myself)
*Listed on SaaSHub: Yes
*Social media following: Very small
*Marketing budget: $0

Is this viral growth? No.

Is it real? Absolutely yes.

14 people signed up for something I built on a broken phone with no budget. That is not nothing. That is everything.

What I Learned

Shipping beats planning

I could have spent six more months planning the perfect app. Instead I spent one month building an imperfect one and launched it. The feedback I got from real users in 11 days was worth more than any amount of planning.

Your constraint is your story

My broken phone became my most powerful marketing asset. People connect with limitation and perseverance far more than they connect with resources and privilege.

If I had built this on a MacBook Pro with a $10,000 budget, nobody would care. Because I built it on a broken phone with $0, people pay attention.

Free tools are genuinely powerful

Firebase, GitHub Pages, Google Gemini - these are not "budget alternatives." They are world-class tools that power real products. The idea that you need to spend money to build something real is simply not true anymore.

Transparency builds trust

Sharing real numbers - including the uncomfortable ones like "0 paid users" - builds more trust than any polished marketing message. People can tell when you are being real with them.

Geography is not destiny

I built this from Jos, Nigeria. Not Silicon Valley. Not London. Not Berlin. Jos.

The internet does not care where you are from. It only cares what you build and how honestly you talk about it.

What is Next

I am focused on:

Growing to 100 users
Getting my first paid user
Launching on Product Hunt
Continuing to share the journey publicly

I do not have a growth hacker. I do not have an investor. I do not have a marketing budget.

I have a story, a working product, and the willingness to show up every day.

That has to be enough. And I believe it is.

To Every Builder Reading This

If you are waiting for:

A better computer
More money
The right time
The perfect idea
Someone's permission

Stop waiting.

The best time to start was last year. The second best time is today.

Your broken tools are not your limitation. They are your story.

Your empty bank account is not your weakness. It is your proof that it can be done without money.

Your geography is not your disadvantage. It is what makes your story unique.

Build the thing. Ship the thing. Talk about the thing honestly.

The rest will follow.

MicroHabit AI is live at (https://microhabitai.github.io/MicroHabit-AI/)

If you are building something with limited resources, I would love to hear your story. Find me on Twitter @MicroHabitAI

Tags: #BuildInPublic #IndieHacker #SaaS #Entrepreneurship #TechInAfrica #Productivity #JavaScript #Firebase