stilodex is a catalog that makes public the visual styles used internally by the artificial intelligence Midjourney to generate images.

In stilodex, each style goes through a process of extraction, curation, classification, and publication, giving shape to an open knowledge base that allows creatives and the curious to understand how AI “thinks” visually.

Why use stilodex

stilodex offers two levels of experience:

1. For enthusiasts from all creative disciplines, it serves as a space for inspiration, reference, and professional growth.
   Create your own collection of generative styles and expand your knowledge by exchanging ideas and opinions with other members of the stilodex community.
   Join as enthusiast »

2. For experts who already work with artificial intelligence or want to take the leap into it.
   In addition to creating your own collection of generative styles and connecting with other users, you can take things to the next level by accessing Midjourney’s --sref codes, which allow you to achieve results with greater speed, efficiency, and control.
   Join as expert » Become an expert »

    

What is Midjourney

Midjourney is one of the most advanced artificial intelligences for image creation. With a surprisingly simple interface for such a highly complex algorithm, it allows anyone to generate professional-level images in seconds. Even so, its use has spread among internationally renowned designers, architects and  artists such as Andrés Reisinger, Hassan Ragab, and Ryan McCoy to explore the boundaries between imagination and creativity through technology.

Unlike other artificial intelligences, Midjourney operates with an internal network containing more than 12 billion visual styles. Each time a user writes a prompt, the system interprets the text, selects, and combines these styles to create an image consistent with the user’s intention. These styles can be unlocked through secret codes in the prompt using a special parameter called --sref.

What is the --sref parameter

When writing a prompt in Midjourney, you can include parameters in the text to communicate directly with the algorithm and fine-tune every detail of the resulting image. For example, --ar defines the aspect ratio, and --chaos controls the degree of freedom you give the algorithm to interpret your prompt.

In the same way, the --sref parameter tells the algorithm which style to apply. By changing the sref code while keeping the same prompt, you’ll obtain stylistic variations of the same idea, while using different ideas with the same sref code will maintain visual consistency across results. This aesthetic coherence is not limited to surface attributes such as light and shadow, color palette, or texture, but also includes abstract and structural concepts that we cannot easily perceive at first glance, such as the interpretation of physical and temporal phenomena, the cultural and historical imagination, and the symbolic and emotional charge, making each code, in itself, a set of characteristics that give life to its own universe.

All images published in stilodex are created with Midjourney, and for that reason, each style in the catalog is a window into a universe contained within itself.