Harrison Hlongwane

# TruthLens AI – Project Description I built **TruthLens AI** to help people check whether the information they read or hear is actually true, instead of just trusting what AI tells them. You can type a claim, use your voice, paste text, upload a document, or provide a website link. TruthLens extracts the claims, searches the live web, and checks them against real sources. **What makes TruthLens different is how it verifies information.** Instead of relying on one AI model to give an answer, I built a system where different AI agents examine both sides of a claim. One looks for supporting evidence, another challenges it, and a final judge makes a decision. The system also: - Reads actual web pages instead of relying on search summaries. - Shows the original evidence, citations, and confidence levels. - Detects and isolates suspicious instructions hidden in websites or documents. - Gives clear verdicts, including Supported, Contradicted, Misleading, Outdated, and Unverifiable. - Shows disagreements and research limitations instead of pretending to know everything. **The goal is to make AI verification more transparent and trustworthy.** TruthLens is designed for students, researchers, journalists, and everyday people who want to check information before believing or sharing it. I wanted to build something that doesn't just give you an AI-generated answer, but actually shows you the evidence behind that answer so you can make your own judgement.