About
The internet is full of clickbait nonsense, especially when it comes to movies and video games. Upcoming releases get flooded with misleading headlines that overpromise and underdeliver—titles that claim big news but offer nothing but speculation and fluff.
"GTA 7 confirmed???"—No, it's just a director casually mentioning the idea. But you had to waste time reading to find that out. Even skimming these articles eats up valuable seconds or minutes.
That's where our app comes in. We use AI to summarize articles and determine whether they’re legit or just clickbait garbage—saving you time and frustration with a quick and accurate verdict.
What it does
Users can paste article links into the web app to generate reports. Once a report is created, they have the option to publish it for others to see. The community reports section allows users to browse and compare reports generated by others.
A feedback system lets users like or dislike the AI’s response, helping improve accuracy. The Chrome extension follows the same mechanism—simply click generate while on a webpage, and it will analyze the content. Users can also view the full report on the web app for a more detailed breakdown.
How We Built It
We started with a web app using ReactJS for the frontend and FastAPI with MongoDB for the backend. To extract website content, we leveraged Beautiful Soup, and for analyzing articles, we used Google Gemini as our LLM.
But why stop there? The goal is to save time, so we extended it with a Chrome extension, allowing users to check articles directly on the websites they visit—making clickbait detection even faster and more convenient.
Challenges we ran into
We had some troubles parsing websites. We initially had a Go server for our scraper but we ran into challenges with anti-scraping systems. So we switched to BeautifulSoup4 in Python since it’s a more mature and simple library for our use case
We also had troubles instructing the LLM to send the proper JSON output, causing weird values to be displayed in both the web app and chrome extension.
Accomplishments that we're proud of
Seeing the app come together. Because we made good progress, the last few hours were spent just polishing, and making both the chrome extension and frontend prettier.
What we learned
Webscraping, frontend development, rest api's. Using LLM'S
What's next for Article Clickbait Detector
We would like to improve our feedback system as it is just a like or dislike mechanism. We're thinking it should be a range such as a 5 star system. We would also like to implement a login system, allowing users to keep track of articles they have generated reports for.
Log in or sign up for Devpost to join the conversation.