🚀 About the Project TruthGuard is an AI-powered platform designed to analyze and verify information shared on Reddit in real time. The goal is to combat misinformation and provide contextual truth scores and insights for trending posts or discussions. Users can get fact-checking summaries, bias detection, and credibility metrics—all while browsing Reddit content.
🌟 Inspiration The idea for TruthGuard was born out of growing concerns around misinformation and biased narratives spreading rapidly on social platforms like Reddit. Traditional fact-checking processes are often manual, slow, and non-scalable. I wanted to build a system that could:
Ingest real-time data,
Use NLP and ML to analyze claims,
Empower users with timely, factual information.
I was also inspired by platforms like NewsGuard, but wanted something more developer-friendly, community-driven, and open-source.
🧠 What I Learned Building TruthGuard taught me a ton about:
🔄 Real-time data ingestion pipelines using Reddit’s API.
🧹 Text preprocessing and enrichment techniques for noisy social media content.
🤖 ML/NLP models for claim detection, sentiment analysis, bias detection, and truth estimation.
🔌 FastAPI for modular backends and websockets for real-time updates.
⚛️ React frontend development with live feed rendering.
🧱 Structuring a scalable backend with modular directory practices.
Built With
- apis
- beautifulsoup4
- cassandra
- cassandra-driver
- core
- db
- embeddings
- fastapi
- handling
- huggingface
- huggingface-hub
- libraries
- mongodb
- natural-language-processing
- nltk
- numpy
- pandas
- processing
- pymongo
- python
- python-dotenv
- requests
- scikit-learn
- scraping
- semantic
- sentence-transformers
- spacy
- testing
- tqdm
- transformers
- uncomment
- uvicorn
- web
Log in or sign up for Devpost to join the conversation.