Expertise
As the holiday season approaches, we wanted to add a special touch to our greetings for customers, partners, and friends. So, this year, we decided to blend technology and tradition by using artificial intelligence to create tailor-made handwritten Christmas cards.
Our journey begins with ChatGPT, an advanced language model that is able to craft unique and personalized messages for each recipient, based on custom hints and a set of information collected in a google sheet. What better way to express our gratitude and well-wishes than with individualized and contextualized messages?
To elevate the personalization even further, we employed another generative AI model to create a custom handwritten card. Thanks to the scientific groundwork by Alexander Graves and the pioneering work on github from sjvasquez, we could repurpose an open-source AI handwriting synthesis algorithm for the crafting of our cards. Each letter is distinct, mirroring the uniqueness of real calligraphy. Our goal was not just to send a holiday greeting but to create a tangible and memorable experience for each person on our list.
Finally, our writing robot takes the stage, seamlessly putting the AI-generated messages and their calligraphy onto paper. The technology serves as a tool, allowing us to scale the personalization without losing the essence of sincerity.
We wrapped everything into a lo-fi style and nostalgic terminal based application built on top of the Python Textual framework. As in the creation of this workflow, we took advantage of many open-source projects; as a token of appreciation, we are giving back the results to the community publishing our results on github. It's a testament to the incredible possibilities that the open-source community provides, allowing us to achieve quickly innovative feats in today's digital landscape. Want to try it? Check out our GitHub Repository.
As the year draws to a close, we extend our warmest wishes to you. May this holiday season be filled with joy, and moments that matter.
Warm regards from The EMBRIO.tech Team