A hybrid fact-checking engine combining the speed of 5 Machine Learning algorithms with real-time web verification powered by the Groq API.
Credible is an advanced fake news detection system designed to combat misinformation. By leveraging a multi-model ensemble of traditional ML algorithms for rapid text analysis and the Groq API for cross-referencing claims against live web data, Credible delivers high-accuracy credibility scores for news articles, tweets, and digital content.
Credible uses a two-step verification pipeline:
- Initial ML Classification: The text is passed through 5 distinct Machine Learning algorithms (e.g., Logistic Regression, Random Forest, SVM, Naive Bayes, Gradient Boosting) to analyze linguistic patterns, emotional tone, and structural red flags associated with misinformation.
- LLM Fact-Checking (Groq API): If the content is flagged as suspicious or needs deeper context, it triggers an ultra-fast web search and evaluation using an LLM hosted on Groq's LPU infrastructure, cross-referencing the claims against trusted online sources in real-time.
- Ensemble ML Detection: Utilizes 5 different machine learning models to ensure robust baseline classification and minimize false positives.
- Lightning-Fast LLM Fact-Checking: Integrated with the Groq API for near-instantaneous natural language processing and web context retrieval.
- Real-Time Web Search: Doesn't just rely on static training data; validates claims against current web events.
- Confidence Scoring: Outputs a clear percentage-based credibility score alongside a brief explanation of why the text was flagged.
- API Ready: Easily integrate Credible's detection pipeline into browser extensions, social media bots, or news aggregator platforms.
- Language: Python
- Machine Learning: Scikit-Learn
- LLM / Inference: [Groq API] (https://groq.com/)
- Web Framework: Streamlit
Follow these instructions to get Credible running on your local machine.
- Python 3.8 or higher
- A Groq API Key
- Clone the repository
git clone [https://github.com/yourusername/credible.git](https://github.com/yourusername/credible.git) cd credible
🔗 Project Link: credible2.streamlit.app