Bubble Secure

Detect passwords in your logs from all sources using Modzy's AI models. Provide multiple sources such as FTP, S3 or SSH and manage it all in one place in just a few clicks.

Comment

Inspiration

Security and AI goes together so naturally because of the large amounts of data needed to be processed. I found that having the ability to scan auth.log and events.evtx files has a huge advantage for internal security audits. Modzy makes this super easy but I wanted to take make it even easier to manage more than one source of logs such as several machines.

What it does

The web application I built allows you to add multiple sources of log files that will be scanned by the Modzy AI models and returned to one location. All the logs will be scanned once a day and the results will show up in one collective dashboard.

How we built it

Using Python Flask and the Modzy API we connected everything together. The infrastructure is on AWS and uses MongoDB as the database. The two scripts, one for scanning and one for checking and updating the results run as cronjobs under crontab.

Challenges we ran into

It was complex to have a situation of uploading so much data to the Modzy API so frequently as well as dealing with different data types being uploaded sometimes in binary and in text.

What we learned

Wow! A lot. There is a lot to learn about the data processing and the pipeline the files need to go through in order to be processed properly. Also setting up and integrating Modzy in the whole process was easy and straight forward and definitely a newly acquired skill.

What's next for Bubble Secure

There are multiple new features I would like to add to platform:

  1. Multiple new sources such as SSH, MySQL, Elastic Search and more
  2. An easier UI for defining the password criteria
  3. An easier way to manage and display the results coming back from the models

Built With

Share this project:

Updates