I'm a Chill Guy: AI-Powered GitHub Profile Roaster
Overview
"I'm a chill guy" is a fun AI-powered web application that analyzes GitHub profiles and delivers witty, humorous roasts based on users' coding habits, project choices, contribution patterns, and profile information. Built with React and powered by Google's Gemini AI, this project combines humor with technical analysis to create an entertaining developer experience.
How It Works
- Simple Input: Users enter their GitHub username
- Profile Analysis: App fetches comprehensive GitHub data via GitHub API
- AI Processing: Gemini AI analyzes the data and generates personalized roasts
- Instant Results: Users receive humorous commentary about their coding journey
- Shareable Content: Roasts can be easily shared with friends and colleagues
What Gets Analyzed
Profile Information
- Bio & Description: Personal information and self-descriptions
- Profile Picture: Avatar choices and profile aesthetics
- Location & Links: Geographic info and social links
- Account Age: How long they've been on GitHub
Contribution Patterns
- Commit Frequency: Daily, weekly, and monthly contribution patterns
- Contribution Graph: The famous green squares analysis
- Streak Analysis: Longest streaks and gaps in activity
- Weekend vs Weekday: Work-life balance in coding
Repository Analysis
- Project Types: Languages, frameworks, and project categories
- Repository Names: Creative (or not so creative) naming conventions
- README Quality: Documentation standards and project descriptions
- Star Counts: Popularity of projects and social proof
Coding Habits
- Language Preferences: Most used programming languages
- Framework Choices: Technology stack preferences
- Project Completion: Finished vs abandoned projects
- Code Quality Indicators: Repository maintenance and updates
Why I Built This
I created this project for several reasons:
- Developer Humor: The tech community loves self-deprecating humor about coding habits
- GitHub API Practice: Wanted to explore comprehensive GitHub data analysis
- AI Integration: Experiment with Google's Gemini AI for creative content generation
- Community Fun: Create something that developers could share and enjoy together
- Quick Project: Build something entertaining in a short timeframe
Tech Stack
Frontend
- React: Modern UI library for interactive user interface
- Tailwind CSS: Utility-first CSS for rapid styling and responsive design
- Vercel: Frontend deployment with global CDN and instant deployments
Backend
- Express.js: Lightweight Node.js server for API endpoints
- Netlify: Backend deployment for serverless function handling
- GitHub API: Comprehensive profile and repository data fetching
- Gemini AI API: Google's AI model for generating roast content
Technical Implementation
GitHub Data Collection
- Profile Fetching: Basic user information, bio, and account details
- Repository Analysis: Public repositories with languages, stars, and update patterns
- Contribution Data: Commit history and contribution graph information
- Rate Limit Handling: Efficient API usage within GitHub's rate limits
AI Roast Generation
- Prompt Engineering: Crafted prompts to generate humorous, not offensive content
- Context Building: Structured data formatting for AI consumption
- Response Processing: Content filtering and formatting for display
- Safety Measures: Content moderation to ensure appropriate humor
User Experience
- Simple Interface: Clean, minimal design focused on the core functionality
- Loading States: Engaging loading animations while processing data
- Error Handling: Graceful handling of invalid usernames or API errors
- Responsive Design: Works seamlessly across desktop and mobile devices
Behind the Scenes
This project taught me that sometimes the best projects are the simplest ones that solve a fun problem. The combination of real data analysis with AI creativity resulted in something that developers genuinely enjoyed using and sharing. It also demonstrated the power of combining multiple APIs to create unique user experiences.
The "chill guy" was a instagram trend, so i thought why not build a website for it.
