Work

ShoeGrab / VyprBot

Snake in the form of letters VB, with pictures of shoes in the background and obfuscated text
C#
WinForms
OCR
Twitter API
OpenCV
Tesseract
Rhino Licensing
Stripe
PHP
Bitcoin
Bitpay

An automated system for RSVP/Online Orders. The application watched Nike's Twitter stream looking for shoe releases and then attempted to use OCR and image manipulation to extract the code required to place an online order. The application would also perform the order for the user.

VyprBot Shop Source | ShoeGrab Source

It was completely new to me, but the culture around Nike shoes in some communities is extremely prolific. As such, people want the newest available and exclusive shoes which leads to a race of who can be the fastest to make an order.

Services to help people buy shoes attract quite some attention, and I thought why not have a fun little project of my own that replicated such a service. That’s where I came up with the VyprBot name (something sinister sounding) where it was hosted at https://vyprbot.com (no longer running) which offered the ShoeGrab client which was designed to listen to announcements on Twitter.

ShoeGrab listened to the live Tweet stream, and relied on running Tesseract/OpenCV/OCR to find the hashtag that was circled in the image that contained the shoe. I trained the application on a few example images, and fine tuned the image handling process to remove as much of the background noise as possible.

The ShoeGrab application looked like: