Inspiration
Fake news is becoming increasingly common due to its sensational and misleading nature. Many people believe misinformation because it appears credible and do not take the time to verify it.
What It Does
FactVision helps users verify the credibility of information with a single click. Using Large Language Models, it assesses the likelihood of a claim being false, provides reasoning, and compiles relevant sources for cross-referencing.
How We Built It
The backend utilizes a Large Language Model API, fine-tuned through prompt engineering for accuracy. Users highlight text, which is sent to the backend for analysis. The AI-generated assessment is then displayed beneath the highlighted content.
Challenges
We faced frontend display issues and bugs in communication between the frontend and backend. Additionally, LLM responses were sometimes unstructured, requiring prompt engineering to ensure stable and correctly formatted outputs.
Accomplishments & Learnings
We successfully built a functional product, learning LLM integration, error handling, Chrome extension development, and teamwork under tight deadlines.
What's Next
Future updates will include user feedback, rating systems for sources, and potentially computer vision to detect fake news in images.
Built With
- api
- css
- deepseek
- html
- javascript
- openrouter
- python
Log in or sign up for Devpost to join the conversation.