Factually

About the Project:

• We built Factually to combat the growing problem of online misinformation. Every day, people scroll through endless news articles, social media posts, and viral claims — but it’s hard to know what’s actually true. Our team wanted to create a tool that makes fact-checking fast, clear, and accessible for everyone.

🔍 Inspiration:

• We noticed how easily false information spreads online and how time-consuming it can be to verify a single statement. During MHacks 2025, we set out to design a lightweight fact-checking assistant powered by AI to help users stay informed in seconds.

⚙️ How We Built It:

• Frontend: Built with React + Vite + Tailwind CSS for a fast, clean, and responsive user interface.
• Backend: Deployed as serverless functions on Vercel, using Node.js for API endpoints.
• AI Core: Integrated Google Gemini models to analyze claims, assign a reliability score, and provide evidence-based reasoning.

Deployment: Hosted on Vercel for easy scalability and instant access.

💡 What We Learned:

• How to integrate cutting-edge AI models (Google Gemini) into a production-ready web app.
• The importance of structuring API prompts to get clear, JSON-based responses.
• Handling CORS, API keys, and deployment challenges while keeping the user experience seamless.

🚧 Challenges We Faced:

• Choosing the right AI model and prompt engineering to get consistent, trustworthy results.
• Debugging API errors caused by mismatched model versions and unsupported runtime configurations.
• Cleaning up auto-generated project boilerplate while customizing the design to match our vision.

Built With

• control:
• frontend:-react-+-vite-+-tailwind-css-backend-&-deployment:-node.js-+-vercel-(serverless-functions)-ai-model:-google-gemini-api-for-fact-checking-and-content-analysis-languages:-javascript
• git
• typescript
• version