💡 Inspiration
Setting up software projects is repetitive. CodePromptLLM turns plain-text prompts into full, working codebases — fast and clean.
⚙️ What it does
It converts a single prompt into a deploy-ready software starter kit: folders, files, routes, auth, DB setup, and docs — all exportable.
🧱 How we built it
- Frontend: React + Tailwind
- AI: GPT-4 via OpenAI API
- Formatter: Custom file structuring
- ZIP Export: JSZip
- Hosting: GitHub Pages / PWA
🧗 Challenges we ran into
- Cleaning LLM outputs
- Handling vague/incomplete prompts
- Keeping folder structure consistent
- Avoiding bloated or broken code
🏆 Accomplishments we're proud of
- Fully working codebase from 1 prompt
- Multi-stack support: MERN, Django, Laravel
- Simple UI, fast ZIP export
- Lightweight and mobile-friendly
📚 What we learned
- LLMs need guided prompt tuning
- Devs prefer clean scaffolds over large templates
- Speed and clarity matter more than features
🔮 What's next for CodePromptLLM
- CLI + Telegram bot version
- Add Swahili/French prompt support
- GitHub auto-push + live deploy
- MPesa, PayPal monetization
- Offline/portable LLM version (Ollama)
Built With
- and
- android
- built-with-**languages**:-javascript
- css
- firebase-hosting-**offline-support**:-pwa-(progressive-web-app)-**authentication-(optional)**:-firebase-auth-**payment-integration-(optional)**:-mpesa-daraja-api
- firefox
- html
- html2pdf-**platform-&-deployment**:-github-pages
- json
- markdown-**frontend-framework**:-react.js
- paypal-rest-api-**database-(optional)**:-supabase-or-localstorage-for-user-sessions-**version-control**:-github-**browser-compatibility**:-fully-optimized-for-chrome
- replit
- tailwind-css-**ai-integration**:-openai-gpt-4-api-**code-structuring-&-parsing**:-custom-prompt-to-code-logic-**packaging**:-jszip-for-dynamic-zip-generation-**export-tools**:-filesaver.js
Log in or sign up for Devpost to join the conversation.