This is a submission for the GitHub Finish-Up-A-Thon Challenge
What I Built
Last year, I walked into a hospital and saw something I still haven't forgotten.
A pregnant woman was being rushed in, bleeding heavily. Everyone was panicking. But what really broke me wasn't the emergency itself; it was what I found out afterwards. She had been having sharp pains for weeks before that moment. Weeks. But her antenatal appointment schedule was finished, and she genuinely didn't know who to reach out to or what to do.
She didn't know if what she was feeling was normal pregnancy discomfort or something serious. So she just... waited. And hoped it would pass. Until one day it didn't.
I was terrified watching that scene unfold, because that woman didn't need a miracle. She just needed someone to tell her, "This is serious, go now."
That's the moment MamaAlert was born in my head.
MamaAlert is an AI-powered maternal health companion built for pregnant women in Nigeria — a country with one of the highest maternal mortality rates in the world. With MamaAlert, any woman can speak her symptoms and get instant AI analysis powered by Gemini with urgency levels — Emergency, Monitor, or Normal.
She can track her pregnancy week, hydration, nutrition, and clinical visits.
She gets culturally relevant food recommendations like ugu leaves, wara, and tiger nuts, foods her body actually knows.
She can find her nearest hospital and call them directly from the app.
Built for Nigerian mothers, in their context, with their foods, for their reality.
Demo
🎥 Demo Video
🔗 Try it live: https://mama-alert-ai.vercel.app/landing
💻 GitHub: https://github.com/Valentinetemi/MamaAlert-AI
The Comeback Story
MamaAlert started at the GNEC Hackathon 2026 Spring — a competition focused on SDG 3: Good Health and Social Good, with prizes including UN-affiliated internships. I was so excited building it. I had the vision, I had the passion, and I genuinely believed in what I was making.
But I ran out of time.
What I submitted at the hackathon was barely alive — a FastAPI Python backend powered by Gemini, and a basic dashboard frontend. The AI wasn't connected to the frontend at all. There was no authentication, no landing page, and no way for a real user to sign up, log in, or experience the product. It existed, but it wasn't real yet.
Life got busy after that, and the repo just sat there incomplete, representing a problem I cared deeply about but hadn't finished solving.
Here's what MamaAlert looked like before — just a dashboard with no way in and a backend that couldn't talk to the frontend
The GitHub Finish-Up-A-Thon gave me the push I needed to go back and finish what I started.
Here's what I fixed and built to bring MamaAlert to life:
Built a full landing page from scratch
Built a signup page with complete Supabase integration to store mothers' data securely
Built a sign-in page with a working authentication flow
Connected the backend and frontend, so the AI triage now works end to end
Deployed the FastAPI backend to Render
Deployed the Next.js frontend to Vercel
Fixed the CORS issue that was blocking AI analysis completely
Fixed Supabase authentication that kept redirecting users to localhost instead of production
Connected all environment variables properly across the frontend and backend
A broken hackathon prototype became a fully connected, deployed, real platform that a pregnant woman in Nigeria can open on her phone today and actually use.
And here's what MamaAlert looks like now
My Experience with GitHub Copilot
Honestly my experience with Copilot on this project was unlike anything I expected.
It didn't just suggest code, it actually connected to my MamaAlert-AI GitHub repository, read through my entire codebase, understood the context of what I was building, and started working. I would give it an instruction and it would plan, build, and push directly to my repo. It felt less like using a tool and more like having a collaborator sitting next to me who actually understood my project.
Here's what Copilot specifically helped me with:
It built my landing page, signup page, and signin page from scratch — creating a new branch called feat/auth-landing, writing all the code including Supabase client configuration, handling the auth flow, and asking for my confirmation before pushing. What would have taken me hours of building and debugging happened in a fraction of the time.
It also caught a backend bug I had missed — a syntax error on line 130 where I was using gemini_data('candidates') with parentheses instead of square brackets. Copilot spotted it, explained exactly why it was breaking the AI response, and gave me the correct fix immediately.
And when my frontend and backend refused to talk to each other because of CORS, Copilot debugged the entire middleware configuration and suggested the exact FastAPI production setup that finally made everything connect.
For a student developer working alone at 3 a.m., trying to finish a product she believes in, that kind of support made a real difference.
This app is for the woman I saw in that hospital last year. And for every woman in Nigeria who has ever been in pain and didn't know who to tell.
Built for Nigerian mothers 🌸
Top comments (0)