The Community
I built Promptlypi for the dreamers. We often use AI in isolation, but I wanted to create a "Social Lab" where your prompt isn't just a command—it’s a conversation starter.
The "Vibe Coding" Journey (and the Struggle)
Let’s be real: this weekend was a battle. I started by trying to hardcode every string, every connection, and every logic gate by myself. But as the deadline loomed, I hit a wall—specifically with Firebase deployment and Base64 image handling.
That’s when I pivoted to Gemini. I shifted my strategy to Vibe Coding: focusing on the architecture and high-level logic while letting Gemini help me crush the bugs and optimize the deployment pipeline. The result? A functional, social AI platform built in record time.
What I Built
Promptlypi is a social ecosystem for AI art:
The Muse Lab: High-fidelity image generation using FLUX.1.
The Community Feed: A place to share art, likes, and comments.
The Remix Engine: A "Vibe Coding" favorite—see a prompt you like? One click loads it back into your lab to remix.
Smart Deployment: Solved the "Secret Key" struggle by implementing a dictionary-mapped environment system on Render.
How I Built It
Backend: Flask (Python)
Frontend: Tailwind CSS (Neo-Brutalist design)
Database: Firebase Realtime DB & Auth
AI: Hugging Face Inference API
The "Secret Sauce": A custom Pillow (PIL) compression pipeline that keeps the community feed lightning-fast by optimizing AI-generated images before they hit the cloud.
Demo & Code
Live Site: https://promptlypi.onrender.com/
GitHub Repo: https://github.com/moksh-jalendra/promptlypi.git
What I Learned
This challenge taught me that being a developer in 2026 isn't just about typing every character; it's about orchestration. By vibe-coding with Gemini, I was able to solve complex deployment errors that would have normally taken days, allowing me to ship a community-focused product in a single weekend.
Top comments (0)