TruthSleuth

Inspiration

Ever thought of how a single viral post can mislead millions before the truth even gets a voice? In an era where misinformation spreads faster than facts, we felt the urgent need for a tool that empowers users to instantly verify claims they come across online. From health myths to political propaganda, TruthSleuth was born to combat misinformation in real-time.

What it does

TruthSleuth is a real-time misinformation detection and debunking platform. Users can input any viral claim — like “AI will replace all doctors by 2030” — and instantly get an Accuracy Score backed by expert opinions, fact-checked sources, and reliable articles. Available as both a browser extension and a web app, TruthSleuth provides transparency, accountability, and peace of mind — right when and where it’s needed.

How we built it

Frontend: Built using React for both the browser extension and web interface. Backend: Powered by a Flask API that handles claim input and processes it via AI. AI Integration: We used the Gemini model to search for verified sources, pull relevant articles, and score the claim based on consistency and expert validation. Data Sources: Integrated APIs from known fact-checking sites and news aggregators (e.g., Google Fact Check Tools, Media Bias/Fact Check). Scoring Engine: A custom logic was designed to assess credibility, source alignment, and claim consistency to generate an Accuracy Score.

Challenges we ran into

Real-time Validation: Ensuring that the verification process is both fast and credible required careful balancing between speed and source depth. Data Credibility: Differentiating between biased sources and reputable ones proved tricky — we had to fine-tune the trustworthiness logic. AI Overload: Gemini occasionally provided generic summaries instead of verifiable claims. We had to iterate on prompt engineering to ensure precise fact retrieval.

Accomplishments that we're proud of

Successfully integrated real-time fact-checking through AI and external APIs. Created a seamless experience across browser and web interfaces. Designed an intuitive Accuracy Score system that helps users understand credibility at a glance. Developed a working MVP within tight time constraints that genuinely addresses a critical global issue.

What we learned

The power of prompt engineering when working with AI models like Gemini. The importance of UI/UX when designing tools meant for quick user decisions. That misinformation isn't just a tech problem — it's a human one, and tech can play a key role in combating it responsibly.

What's next for TruthSleuth

Crowdsourced Reporting: Enable users to flag misinformation and contribute fact-checks. Multilingual Support: Expand beyond English to tackle global misinformation. Mobile App: Build a companion mobile version for on-the-go verification. Partnerships: Collaborate with news outlets, educators, and social platforms for broader impact. AI Improvements: Fine-tune Gemini for better context understanding and precision.