Contendator is an application designed to detect, filter, and revise racist, stereotypical, and prejudiced comments related to specific religions, castes, colors, or beliefs. This project was created by Govind, Srichand, Marrea, and Alna and won 3rd place at the Amaze Hackathon held at GEC Kannur.
Some of Contendator's main features include:
- Content Filtering: Upload content, and the app will identify and remove problematic language, replacing it with an appropriate alternative.
- Multiple Input Formats Supported:
- PDF Files: Analyze text content from PDF files.
- MP4 Files: Process audio or video in MP4 format using speech-to-text conversion for analysis.
Technologies used in this project:
- Flask – For handling backend requests.
- Node.js & Express – For efficient server-side operations.
- OpenAI’s GPT-3.5 Turbo – Generates non-offensive alternatives to flagged content.
- Whisper – Converts speech to text in MP4 files for content analysis.
Planned improvements include:
- Enhanced UI for easier content management.
- Support for additional file formats and languages.
- Advanced customization options for filtering sensitivity.
Contributions are welcome! If you’d like to suggest improvements or find any issues, please feel free to open an issue or submit a pull request.