Verifact

Inspiration

We are constantly surrounded by news on social media, messaging apps, and short videos. A lot of this information spreads very fast, but it is often unclear whether it is true or misleading. During sensitive situations like conflicts, market crashes, or emergencies, false or incomplete news can create panic and confusion. This problem inspired us to build Verifact, a platform that helps people check the authenticity of news before believing or sharing it.

What it does

Verifact helps users verify the authenticity of news before believing or sharing it. Users can paste news text or upload screenshots, and the system analyzes the content using AI to provide a trust score, a clear explanation, and possible sources. It also allows users to ask follow-up questions about a news article and receive concise, context-aware answers. Additionally, Verifact offers a personalized news feed based on user interests, helping users stay informed while reducing exposure to misleading or unreliable information.

How we built it

Verifact was built as a full-stack web application. The frontend is developed using Next.js to provide a fast and smooth user experience. The backend handles news verification, article-based question answering, and personalized news feeds. We integrated Gemini AI models to analyze news text and images and provide a trust score with explanations. Real news is fetched using production-safe APIs like GNews and Google News RSS. To improve reliability, we implemented caching and fallback mechanisms so the application continues to work even when API limits are reached.

Challenges we ran into

We faced several real-world challenges during development. Some news APIs restrict usage in production environments, which required us to explore alternative providers. AI rate limits forced us to optimize when and how verification requests are made. We also had to ensure that uncertain or unverifiable information is clearly labeled instead of being presented as fact. Supporting both text and image-based news while keeping consistent results was another challenge.

Accomplishments that we're proud of

We successfully built a working, end-to-end news verification platform that analyzes real news content using AI. Verifact can verify text and image-based news, provide trust scores with explanations, answer user questions about articles, and deliver personalized news feeds. Despite limitations like API restrictions and AI rate limits, we implemented caching, fallbacks, and production-safe news sources to ensure the platform works reliably in real deployment environments.

What we learned

While building Verifact, we learned how important system design is when working with AI services. Free-tier AI models have strict rate limits, so caching and selective verification are essential. We also learned the value of prompt design, clear explanations, and transparency when presenting AI-generated results. Beyond technical skills, we gained a better understanding of how trust and reliability matter in user-facing applications.

What's next for Verifact

The next steps for Verifact focus on making the platform more powerful, scalable, and widely usable. We plan to add video analysis to detect deepfakes and misleading clips, improve personalization using user behavior, and integrate more trusted news sources for stronger verification. We also aim to optimize AI usage with better caching and scalable infrastructure, and expand Verifact beyond the web through browser extensions and mobile apps so users can verify news directly where they consume it.

Built With

• g.news
• nextjs
• postgresql
• supabase
• tailwind
• typescript