Clarif AI
AI-powered misinformation detection for social media.
The Problem
Social media is flooded with misleading posts, AI-generated fakes, and viral misinformation. Before this project, there was no tool for Nepali users to quickly check the trustworthiness of a Facebook post or shared image in real time.
What They Built
Clarif AI is a mobile app that lets users share any post, text, or image directly from Facebook or other platforms. The app sends both the text and image to a custom backend, which runs advanced AI analysis: it checks for AI-generated content, fact-checks claims, and analyzes images for authenticity. The backend returns a clear trust score, a color-coded verdict, and a list of red flags—so users can instantly see if a post is likely real, misleading, or fake. The integration required deep changes to both the Flutter frontend and FastAPI backend, wiring up real AI pipelines, robust error handling, and a new response format that the app could display in a user-friendly way.
Impact
Clarif AI became a hackathon project to offer end-to-end, real-time misinformation analysis for social media. The team delivered a working demo that handled text, images, and even full Facebook shares, with a robust backend that gracefully handled errors and missing dependencies. The project set a new bar for technical depth and practical impact at SXC Sandbox.