AI Image Generation Workflows

This project is an ongoing exploration of AI image generation workflows built in ComfyUI. It is less about producing a finished product and more about researching how these systems work in practice, how they can be controlled, and where they are most useful in a real workflow.

A large part of the work has focused on txt2img and img2img pipelines, along with understanding how to push them toward more reliable and usable results. One of the main areas of interest is character consistency, since that remains one of the weaker parts of many AI image-generation systems. The goal is to better understand how to keep a recognisable visual identity across multiple outputs rather than treating each image as a one-off result.

This is still an ongoing area of experimentation, so the direction will continue to evolve. The work will likely expand into more refined workflow setups, stronger consistency methods, and a better understanding of how these tools can support real design, product, and visual-development work rather than just producing isolated images.