Plagia

Inspiration

I recently went to two hackathons at Stony Brook University. At the second one, I saw a team using my team's project idea from the first hackathon, without giving us any credit! They basically copied our whole idea and when I tried to report to the organizer, they told me "You can't prove it is your idea because it is a very common app"

However, at HackPSU, I saw a fantastic project by Kartikey Pandey: DevSpot by Kartikey Pandey (you should definitely check it out!). It sparked an idea for me to add new features to something similar, but in a way that respects the original work.

What it does

Plagia is a plagiarism detection software for hackers which helps ensure credibility and originality in works, hence promoting safe coding for developers. It has a taipy link for an assistant that gives feedback to a project, and also it uses Verbwire to make a smart contract to own the project using your wallet made easy with MagicPlot.ai and Verbwire

How we built it

The backend uses Python to scrape Devpost. This allows Plagia to gather information about projects submitted to various hackathons. Then we have an AI model trained to compare projects. When you provide Plagia with a project URL and a reference hackathon URL, the AI compares the project to others submitted in that hackathon. This comparison helps identify similarities and potential plagiarism. The frontend, built with Next JS, is the user interface you interact with. It allows you to input project and hackathon URLs, analyze the project, and see the AI's comparison results. The transaction part was creaeted using MagicPlot.ai and Verbwire

An additional feature is to connect to a wallet that will allow user to make a smart contract to own the project if it matches all the requirements.

Challenges we ran into

At first the idea was to compare the project to every single previous hackathons projects, but I realized that it was too much data and Python was running very slow to loop over every single projects and run the AI to compare it to the selected project. And also, tried to scrape MLH websites to get all past events but since the elements in the DOM are rendered dynamically, I couldn't access them the way I wanted to, I had to log all the past events from the browser console. But then again realizing later that it is going to take a lot of time during the demo waiting for the python to execute them all, so I used that bug to be a new feature.

Accomplishments that we're proud of

The app is running as expected

What we learned

Python is indeed a very slow programming language.

What's next for Plagia

If I had more time, I would have added more features to it such as comparing it to the winner projects only, to projects in different locations only and so on.

Built With

• magicpolt.ai
• verbwire
• web-scrapping