niche.files scans your codebase to find any security vulnerabilities and then reports them back to you via subway surfers brainrot. we make it easier to find and understand your vulnerabilities and how to fix them.
How we built it
we built it using next js fullstack. after the user inputs the codebase, we use typescript to break down the code, find vulnerabilities, and write a script for the brainrot video. we use the elevenlabs api to take the script and convert it into audio that can play along the video.
Workflow
1) Codebase is scanned Analyzes a GitHub repo and identifies bugs, bad practices, or security issues. The findings are written into a plain .txt script.
2) Text to spoken explanation Takes .txt file and converts it into high-energy, angry, comedic speech using ElevenLabs.
3) Spoken explanation to visual content The generated audio is placed on top of a pre-recorded Subway Surfers video, creating an immediately engaging visual context.
4) Single shareable artifact is produced Outputs a single MP4 video that explains what’s wrong with the code and is optimized for attention and sharing.
Challenges we ran into
at first, we thought we were allowed to use the minimax video api, so we were working on a similar but different idea that focused on creating ai brainrot videos to teach developers common issues about programming. we didn't realize that we didn't have access to their video api until more than 2 hours into the hackathon (40% of the way in). we had to completely pivot after wasting time
we also ran into problems with moviepy and burning the subtitles into the video, at the end we removed the subtitles as it was causing issues.
Built With
- elevenlabs
- minimax
- moviepy
- python
- typescript
Log in or sign up for Devpost to join the conversation.