"You see how this looks, right?" - Chris Hansen, Dateline NBC
redflag is a lightweight local tool for analyzing conversations and transcripts to detect suspicious or predatory behavior. It is designed to help flag patterns that may indicate grooming, manipulation, or predatory intent so users can review risky interactions more quickly.
-
Install the required model files: python src/models/get_models.py
-
Update the transcript path in main.py if needed.
-
Run the app: python main.py