RealFeed: AI-Powered News Verification

🎯 The Problem

Misinformation spreads rapidly online, making it difficult for people to distinguish credible news from fake content.

💡 The Solution

RealFeed uses AI to analyze news headlines and provide instant credibility assessments with confidence scores.

🛠️ Technical Stack

• Backend: Flask with RESTful API
  
• AI: Hugging Face Inference API
  
• Data: NewsAPI for real-time headlines
  
• Frontend: Modern glassmorphism UI
  
• Core Logic: Custom AI pipeline + fallback classifier
  

🔧 Core Features

• Instant headline verification
  
• Confidence scoring (0–100%)
  
• Real news integration
  
• Topic-based search
  
• Mobile-responsive design
  
• Dual-mode classification (primary AI + fallback model)

🚧 Challenges Overcome

• Model Reliability: Tested multiple Hugging Face models to find the most stable and accurate one.
  
• API Changes: Adjusted to endpoint updates and model availability issues.
  
• Accuracy: Improved from sentiment-based checks to dedicated fake-news detection models.

📈 Outcome

A reliable AI-powered tool that helps users quickly evaluate the credibility of news headlines using automated analysis.

🔗 Live Demo and Source Code

• Live Application: https://realfeed-ai.onrender.com
  
• Source Code: https://github.com/hub-CODE/RealFeed-ai

Built With

• bert-architecture
• css-grid
• css3
• flask
• flexbox
• git
• github
• html5
• hugging-face-inference-api
• javascript
• natural-language-processing
• newsapi
• probability
• python
• render
• restful-apis