This is a submission for the GitHub Copilot CLI Challenge
What I Built
🛑 The Problem: I Was Drowning in "Productivity"
It was 2 AM. My desk was a mess.
I had 14 browser tabs open. A Notion doc I hadn't touched in weeks. A Todoist list that was mocking me. A goal I set in January that was already dying quietly.
I was drowning in tools that were supposed to save me. I realized I didn't need another "to-do list." I didn't need a static database.
I needed a brain.
So, I opened my terminal, cleared the screen, and typed:
gh copilot suggest "How do I build a life operating system where tasks, habits, and goals share a single consciousness?"
7 days later, I shipped CorteXia.
🚀 What is CorteXia?
CorteXia isn't just an app; it is an AI-Powered Life Operating System. It combines the flexibility of Notion, the visual beauty of Linear, and the intelligence of a personal assistant into one unified workspace.
But here is the killer feature: It talks back.
💎 The "Superpowers"
- 🗣️ Voice-First Control: I used the Web Speech API so you can talk to your dashboard. "Hey, I just spent $15 on coffee." Boom. Budget updated.
- 🧠 Persistent AI Memory: It doesn't just answer; it remembers. It knows your name, your deadlines, and that you're trying to quit sugar.
- 🔥 The "GitHub" for Habits: A 365-day contribution graph for your life habits. Because if it's not green, did it even happen?
- 🕸️ The Neural Network: Tasks are linked to Goals. Expenses are linked to Budgets. Moods are linked to Productivity. Everything is connected.
👇 Give it a Try!
I poured my heart (and a lot of coffee) into this.
(P.S. Try the Voice Mode. It feels like magic. ✨)
My Experience with GitHub Copilot CLI
I’m Lokesh A, a third-year Computer Science Engineering student focused on full-stack development.Building real time projects!
This wasn’t just CRUD. It was a Next.js 16 + MongoDB (12 Collections) + AI-powered system with deeply connected logic.
GitHub Copilot CLI became my force multiplier.
Architecture
Designing a high-performance, bidirectional schema for Tasks, Goals, and Habits required system-level thinking.
gh copilot suggest "Design a Mongoose schema optimized for dashboard-heavy reads with bidirectional references"
It didn’t just generate code — it explained embedding vs referencing trade-offs and helped me make better architectural decisions.
AI Context Strategy
To avoid token overflow while injecting live user data into the LLM:
gh copilot explain "Inject relevant user data into an LLM without exceeding token limits"
It guided me toward a selective context strategy (today’s tasks + active goals only), which shaped the AI design.
Debugging
When optimistic UI updates caused duplicate entries:
gh copilot explain "React optimistic update causing duplicate keys when toggling quickly"
It identified the need for debouncing and helped structure a clean solution.
Copilot CLI didn’t replace my learning.
It accelerated it.
It helped me think bigger, ship faster, and build with more confidence.
🛠️ The Tech Stack (Vibe Coding Edition)
- Framework: Next.js 16 (App Router + Turbopack)
- Language: TypeScript 5 (Strict Mode)
- Database: MongoDB Atlas (12 Collections)
- Styling: Tailwind CSS + Framer Motion (for those buttery smooth 60fps animations)
- AI: Multi-LLM Fallback System (Groq + OpenAI)
- Voice: Native Web Speech API
🔮 The Bigger Picture
We talk a lot about AI replacing developers. After building CorteXia, I believe the opposite.
Copilot CLI didn't replace my creativity; it accelerated it. It handled the boilerplate, the schema validation, and the obscure regex patterns, so I could focus on User Experience and Logic.
CorteXia went from a "weekend idea" to a production-grade application in record time.
That's not AI replacing a developer.
That's a developer with a superpower.
https://dev.to/itslokeshx/i-built-a-second-brain-that-actually-thinks-with-github-copilot-cli-5946
Lokesh