zylte

Your proactive AI co-founder — it understands your life, structures your goals, simulates your future, and accelerates your execution

Comment

Project Story: ZYLTE — AI Co-Founder

Inspiration & Vision

ZYLTE was born from a simple but deep frustration: why do dozens of tools exist — Notion, Trello, ChatGPT, LinkedIn, Fiverr — yet none of them truly understand who you are and act accordingly?

As a student in Applied Economics (L3), an AI developer, and an entrepreneur based in Cotonou, Benin, I lived this contradiction every day. I had clear goals and ambitious projects, but no system understood me well enough to help me execute them intelligently.

The core vision was to move from a passive AI — one that waits for commands — to a proactive AI that anticipates the entrepreneur's needs, analyzes risks in real time, and proposes concrete execution plans. ZYLTE was born from one question: what if your app learned who you are, anticipated your needs, and acted before you even asked?

What I Learned

Building ZYLTE was a deep dive into complex Edge Function architecture and asynchronous state management. I learned the critical importance of synchronization between persistent storage (Supabase) and the real-time user interface.

I also learned that the true value of an AI product lies not in its features, but in its depth of understanding of the user's real life. Every piece of data collected must serve a decision — every user action must enrich their digital profile.

The XP progression system follows a logarithmic logic to sustain long-term engagement:

$$XP_{next} = XP_{base} \times \log_{10}(Level + 1)$$

How It Was Built

The project was built entirely on MeDo, structured around a modern and robust technology stack, with an architecture organized into 6 core tabs: Home, Execution, Simulation, Network, Finance, Profile — featuring an adaptive onboarding flow whose questions vary based on the detected profile (student, entrepreneur, creator, freelancer, or multi-profile combinations).

⚡ Challenges Faced

The path was far from smooth. The most frustrating difficulty: applied corrections were not always taken into account immediately, requiring repeated iterations and extraordinary patience. Every detail needed precise refinement before changes became truly effective.

Among the most critical bugs: an onboarding screen stuck in an infinite loading loop blocking access to the entire app, corrections that applied only partially, and sessions that reset the user profile on every reconnection. These grueling iteration cycles consumed a significant number of credits, adding extra pressure throughout development.

What Impressed Me Most

The feature that changed everything: the AI Test Engineer. Its ability to analyze the entire codebase, test every action on every page with surgical precision, and detect invisible anomalies was a true game changer. The corrections it proposed not only resolved critical bugs but elevated the overall quality of the project to a new level.

$$Efficacité = \frac{Successful_Actions}{Total_Tests} \times 100\%$$

ZYLTE — Your vision, our execution.

Built With

Share this project:

Updates

Submission history