💡 Inspiration

The use of QR Codes are rising exponentially. QR Codes are used for transactions, forms, and lots of other things. Countries such as Indonesia, UK, New Zealand, and more also scan QR Codes anytime they enter a public space as a way to trace for covid. However, this creates a huge problem in security, where malicious QR Codes can affect huge amounts of people easily.

Even the Federal Bureau of Investigation (FBI) warned Americans two months ago that cybercriminals are using maliciously crafted Quick Response (QR) codes to steal their credentials and financial info.

To solve this ongoing issue, we created SafeQR.

💻 What it does

SafeQR uses a triple check method to ensure your QR code is not malicious. We check for redirects, viruses, databases of phishing website, and more! You can also few the WhoIS data of the QR website as well, so you are always sure who owns the website you’re visiting.

🔨 How we built it

We hosted a database on MongoDB and a REST API on Google cloud functions. The REST API provided functionality to the front end and utilized API calls as well as data retrieval from lists of well known URLS and the output of the below mentioned ML approaches. We investigated several well known approaches to malicious URL detection. We based our approach on the following resources: 1: https://towardsdatascience.com/predicting-the-maliciousness-of-urls-24e12067be5 2: https://www.kdnuggets.com/2016/10/machine-learning-detect-malicious-urls.html

After using these to curate a list of URLs, we further used the VirusTotal API to give us a confidence score of the URL classification.

We use the mongodb to then cache the results. So subsequent calls are faster and provide better latency for other users



Built With

Share this project: