Readist is our take on two challenges for this edition of HackUPC:
First of all, as we proved last season we love computer vision so we took the Itequia challenge in an attempt to keep exploring this world (this time with machine learning an very powerful tools!). But as it stands pure text transcription is a bit raw for the end user, that's why we've decided to combine this tech with Twist. A tool that turns blackboard and sketches into edible rich text fits perfectly in an environment designed for remote workers and decentralized businesses.
So, what did we do?
We have used the Microsoft Azure, a powerful cloud platform with neural network that recognizes text in an image. Our task was deploying that software and making the text edible and with easy access from an html document. We have to say that this is a very powerful tool and it made itself easy to pilot, 10/10 would code again!
On the other hand, the Twist implementation was a huge challenge. Without any knowledge on web apps or back end we tried to set up from the ground an Amazon Web Service Lambda Function (a serverless solution for a web app). All of that to make us able to, through the Twist Zappier integration, fetch your images in a thread on any channel and send them to the cloud to come back as clean, formatted (and formattable) text.
And of course, the glue that keeps this pieces together our beloved Python.
We learned a ton during this hard project and although we know the boons of machine learning aren't perfect (yet) we sincerely hope you enjoy it!