RepoRadar is an AI-powered tool designed to automatically analyze GitHub repositories for bugs, security vulnerabilities, code smells, and potential improvements.
Inspiration
The idea came from observing how developers, especially in hackathons or fast-paced projects, often struggle to quickly assess code quality. Manually reviewing repositories is time-consuming and error-prone. I wanted to build a tool that could provide actionable insights instantly using AI.
What I Learned
Through this project, I gained hands-on experience in:
Integrating AI models with code analysis pipelines.
Handling multi-modal inputs, such as Python scripts and Jupyter notebooks.
Designing a dynamic and interactive web dashboard for displaying reports.
Using environment variables and secure API management best practices.
I also learned the importance of error handling, memory optimization for file uploads, and how to provide meaningful metrics to guide developers.
How I Built It
The project uses a three-tier architecture:
Frontend (Presentation Layer) – Built with HTML, CSS, including collapsible dashboards for metrics like Bugs, Security, Code Smells, and Improvements.
Backend (Application Layer) – Python,FastAPI interacts with AI APIs, and generates structured reports.
AI Layer – AI models analyze code to identify issues, suggest improvements, and generate summaries. Note: Right now, the tool only analyzes Python (.py) files and Jupyter notebooks (.ipynb). Other file types are ignored.
Challenges
API Security: Storing and using API keys securely without exposing them.
File Handling: Large files required efficient streaming instead of reading into memory all at once.
Structured Reporting: Displaying metrics in a clean, interactive way while keeping the dashboard responsive.
Log in or sign up for Devpost to join the conversation.