- Published on
Your Song Zest
- Authors
- Name
- Jorge Zozaya
We are fortunate to live in a unique moment where we can train models for very specific tasks and then orchestrate them with algorithms that respond to communication strategies.
Last year was full of challenges and great discoveries. Mid-year, I became a ghost producer, and to achieve this, we had to develop an unparalleled content pipeline. This pipeline allowed us to discover the possibilities of content capital, enabling us to generate and use music and images to create customized soundtracks.
Imagine having an inexhaustible and at the same time unrepeatable source of content that captures the unique essence of the brand, all while maintaining a visual standard that ensures consistency and quality.
1- Multi step prompt-building website
2- ML model that takes the prompts and generates a piece of music
3- ML model that generates an image based on the prompt
4- Algo that bakes both assets into a video for the user
