Phishy

• 

  Phishy Logo

• 

  Phishy Functionality

• 

  Training the ML model

Inspiration

Every year, millions of people are targeted by phishing scams. Raymon, a member of our group, shared that his grandparents had even been targeted by phishing attacks in the past, something that other group members could relate to. We were inspired by this, and our aim was to help these individuals by providing effective protection against these cyber threats. Through diligent collaboration, we successfully created a powerful tool that enhances online security and offers proactive defense. With our solution, Raymon's grandparents and many others can confidently navigate the digital landscape, free from the worries of phishing attacks.

What it does

Our solution is equipped and trained AI technology, enabling it to effectively detect and prevent phishing attacks within your Gmail. It automatically reads the phishing emails, and highlights them in red to alert you.

How we built it

We used Python and the TensorFlow library to train the AI model, then we created a Google extension to scrape through Gmail.com, connecting the two using a Google Cloud server.

Challenges we ran into

1. There was an issue with web scraping and getting all the emails
2. Issue with Chrome Tabs API- couldn't inject script into the webpage
3. TensorFlow.js - couldn't import and link it to the extension
4. Yarn wouldn't download on a windows computer
5. Getting the Flask API to work properly on Google Cloud

Accomplishments that we're proud of

1. We completed the website
2. Extension can scrape through all the emails
3. Overcame TensorFlow challenges

What we learned

1. HTML and CSS is easier to learn than you think
2. How to make an extension
3. You will not get any sleep
4. How to link many things together from different servers
5. How to train the model using datasets from the internet
6. How to build a Flask API

What's next for Phishy

1. Connect it to a server and get TensorFlow directly to the extension to optimize speed
2. Create an account system with Auth0, so users settings sync across their devices.

Built With

• chromeapi
• chrometabsapi
• css
• gmail
• html
• javascript
• python
• replit
• tensorflow
• vscode