An image and video lab using ComfyUI, Windows with an NVIDIA GPU and Mac with Apple Silicon.
I configured the local environments and workflows for generation, editing, background removal and upscaling. I use models such as SDXL and Juggernaut XL, alongside cloud services as needed.
I keep prompts, references, parameters and files from each production to review results and resume work. Visual references help maintain character continuity across images.
The lab is used to prepare material for campaigns, courses and product demonstrations.
ComfyUI, image models and Python organization
ComfyUI organizes generation through node-based workflows. SDXL and Juggernaut XL are among the models used, with local execution in compatible environments and external services depending on the production. Python complements work preparation and organization; not every model runs the same way on Mac and Windows.
What this development contributes
Saving references, prompts and parameters makes it possible to revisit an image and understand how it was produced. Final selection requires visual review and suitability for use: campaign artwork has different requirements from interface assets. This is generative AI integration in production, not training a proprietary foundation model.
