Inspiration
Many smaller organizations such as startups tend to have less refined code reviews due to a lack of funding or other resources
What it does
CodeGenius uses a specially engineered GPT-4 prompt to analyze submitted code for vulnerabilities, and provides possible solutions for them
How we built it
The app is hosted using Google Cloud Compute Engine with the frontend written in React.js and TypeScript, with libraries like AntDesign, TailwindCSS, and Code Mirror, and the backend built with Flask.
Challenges we ran into
We ran into many technical challenges. Seth had to figure out a whole new cloud platform. However, his AWS experience came in handy despite the differences between the two platforms. The OpenAI library also does not have very good documentation.
Accomplishments that we're proud of
The program was very technically difficult to develop, and we're very proud to have pulled it off. Developing a fully-functioning app of this caliber is a feat we're proud to have done
What we learned
Seth learned a lot about GPT-4 and Google Cloud, and Chau learned about RESTful APIs and microservices.
What's next for CodeGenius
We would like to improve the UI of the app and add more functionalities to the project such as limiting file size, more structured output, etc.
Built With
- ant-design
- codemirror
- flask
- google-cloud
- gpt-4
- python
- react
- tailwindcss
- typescript
Log in or sign up for Devpost to join the conversation.