A platform for detecting, analysing, and responding to harmful/hateful content targeting women on social media (Twitter/X).
redpill-radar/
├── analyse/ Content analysis & moderation API (Python/FastAPI + Groq LLM)
├── collect/ Twitter/X crawler agent that gathers content for analysis
├── rebutt/ Rebuttal generation module
├── data/ Shared SQLite database (used by analyse and rebutt)
└── venv/ Shared Python virtual environment
FastAPI backend that ingests tweets, classifies them using Groq LLM (safe/harmful, age category, abuse subcategories), and tracks processing state in SQLite. See analyse/README.md for full API docs, setup, and database schema.
Frontend agent responsible for searching/crawling Twitter/X to gather content that may contain harmful or hateful speech targeting women. Feeds content into the analyse API.
Module for generating rebuttals or counter-narratives to flagged harmful content.
# 1. Set up the virtual environment (from repo root)
python -m venv venv
venv\Scripts\activate # Windows
# source venv/bin/activate # macOS/Linux
# 2. Install dependencies for the analyse module
pip install -r analyse/requirements.txt
# 3. Configure environment
copy analyse\.env.example analyse\.env
# Edit analyse/.env with your Groq API key
# 4. Run the analysis API
cd analyse
uvicorn app.main:app --reload --host 0.0.0.0 --port 8000API docs available at http://localhost:8000/docs once running.