💡 We're looking for your creative genius to level up the Penpot challenge.
Share your groundbreaking ideas and let's redefine what's possible here ! 🔥
As a technologist, I want to generate a set of coherent raster images and texts that can fill in selected placeholders in a Penpot Board structure found in a Penpot serialized output so I can speed up a mockup process
Approach #1 | Text-to-Text [More]
The system will help users to generate text for one or multiple placeholders depending on the element context.
Approach #2 | Text-to-image or Image-to-text in whole design [More]
The system will help users to generate raster images to fulfill the design based on the existing text in the design. The text/image could be mutually replaced.
[A1] Text-to-Text using a language model appears more feasible in the short term. On the other hand, [A2] Text-to-image or Image-to-text in whole design could potentially demonstrate higher effectiveness for the clients; however, the implementation process is more complex with moderate risks.