# Pixel Interfaces: AI Workflows in ComfyUI > Many currently reach for a hosted service, sending a prompt to a remote server through an interface designed by someone else, and wait for a result, carrying costs that are easy to ignore: energy, latency, dependency, and expense. Local generation inverts that relationship. It i… ## At a glance - Format: On-demand course - Price: €35.00 (list €50.00) - Difficulty: Beginner - Duration: 6 Hours - Schedule: Aug 1-2, 2026 - Instructors: James McBennett - Categories: Artificial Intelligence - Software: ComfyUI, Python, Gradio - Students: 83 - Rating: 4.0 / 5 - Canonical: https://paacademy.com/course/pixel-interfaces-ai-workflows-in-comfyui - Enroll: https://paacademy.com/course/pixel-interfaces-ai-workflows-in-comfyui --- ## About this course Pixel Interfaces: AI Workflows in ComfyUI is a course that teaches designers to build controllable, local AI image generation workflows in ComfyUI and turn them into interactive design interfaces. Across a two-day program, students construct node-based text-to-image and image-to-image pipelines using ControlNet, LoRA, and depth-map conditioning, then package a chosen workflow into a shareable, browser-based interface using Python, Gradio, HTML, CSS, and JavaScript. The course is Beginner level and results in a complete, deployable interactive tool such as a material explorer, façade variation generator, or comic-layout design interface. ## What you'll learn - Understand core principles of diffusion-based generative models and how they produce images - Navigate and operate ComfyUI for generative workflows - Build, modify, and extend pipelines for image synthesis and image-to-image translation - Apply conditioning methods such as ControlNet, depth maps, and style inputs to guide outputs - Critically evaluate outputs in relation to design intent and project goals - Package workflows into shareable, interactive interfaces ## Methodology The workshop is organized around the construction of a deployed, interactive interface capable of transforming between text prompts and pixel images. Inspired by a past student project by Christina Christoforou and Renuka Deshpande in 2025, this workshop builds an interface that generates comic book layouts, turning a few text and image inputs into a visual narrative of a project. Archigram pioneered the use of comic layouts in the 1960s to propose radical urban futures such as Plug-In City and Walking City. OMA used sequential narrative drawing to move a reader through the program and circulation in ways that plans and sections cannot. BIG extended this with "Yes is More" in 2009, using comic-style storyboards to make complex ideas legible, showing not just what a building looks like but also why it exists. This workshop builds on that history. Working in ComfyUI, participants will build and modify node-based pipelines for text-to-image generation and image-to-image transformation that use architectural drawings, reference images, and depth maps as inputs. Its graph-based structure is similar to that of Grasshopper 3D, making the underlying logic of each workflow visible and modifiable, which suits both learning and iterative experimentation. The workshop will move across model types, examining the differences between base models, fine-tuned models, and LoRA adaptations trained on specific architectural styles or practices. The final deliverable is a working interface that accepts one or more input types and returns meaningful generative output. A pipeline is not complete until someone else can use it and, in using it, reveal what the next iteration might become. As Don Norman observed, the moment you put something in front of users, you discover everything you got wrong. The brief is open and not limited to comics and stories: any interface that takes architectural inputs and returns generative output is in scope. A material explorer, a facade variation tool, a site atmosphere generator, and a prompt library with live preview. The tool should be legible enough for someone else to use it without explanation. The workshop progresses through three connected phases: first, participants build foundational skills in ComfyUI by experimenting with structured text-to-image and image-to-image workflows, focusing on how variables like models, samplers, guidance scale, and step count affect outputs. The second phase expands to more advanced workflows using ControlNet and LoRA, in which participants modify and intentionally deconstruct node-based systems to understand how different configurations affect results and their suitability for specific tasks. In the final phase, participants will design and build their own browser-based interface using HTML, CSS, JavaScript, and Python to make a chosen workflow accessible, resulting in a unique tool accompanied by documentation and a reflection on their design decisions. ### Program: **Day 1: Foundations – Image Generation and Control** - Introduction to diffusion models and their role in generative design - Set up and first experiments in ComfyUI with text-to-image workflows - Image-to-image and ControlNet workflows using architectural references - Exploration of models, LoRAs, and how parameters affect output quality and cost - Structured comparisons and group review of generated results **Day 2: Extension – Interface Building and Deployment** - Introduction to browser-based generative interfaces and reference examples - Selection of a Day 1 workflow to develop further - Building interactive tools using HTML, CSS, JavaScript, Python, and Gradio - Packaging workflows into shareable, user-facing interfaces - Final presentations and critique of completed interfaces ## Curriculum ### Session 1 - Introduction & PAACADEMY Updates (06:06) - Introduction To Local AI (53:45) - ComfyUI Subgraphs And ControlNet (54:31) - Image ControlNet And Outpainting (54:04) ### Session 2 - Fine-Tuning LoRAs (55:53) - Gradio Interface Testing (01:00:44) - Gradio Integration (24:20) - Architectural Comics Workflow Automation (24:14) ## FAQ ### Do I need prior experience with ComfyUI or AI image generation? No. The workshop introduces the fundamentals of diffusion models and ComfyUI workflows from the beginning. Basic familiarity with computational design concepts is helpful but not required. ### Will I learn how to create my own AI image-generation workflows? Yes. You will learn how to build, modify, and extend ComfyUI pipelines using various models, parameters, and conditioning methods, rather than relying on pre-made workflows. ### What type of projects can I create using these workflows? The workshop explores architectural applications such as design visualization tools, material explorers, facade variation generators, site atmosphere studies, and other interactive generative design interfaces. ### Do I need programming experience to participate? No programming experience is required. The workshop introduces the necessary concepts and demonstrates how Python and web-based tools can be used to extend and share generative AI workflows. ### Will I learn how to control AI-generated results? You will explore methods such as ControlNet, depth maps, reference images, and style inputs to guide AI outputs according to specific design intentions. ### What is ControlNet and how does it guide AI image generation in ComfyUI? ControlNet is a conditioning method that adds spatial or structural control to diffusion-based image generation by using inputs such as depth maps, edge maps, pose skeletons, or line drawings to guide where and how the model generates content. Instead of relying entirely on a text prompt, ControlNet lets you use an architectural drawing, a depth image from a 3D model, or a reference photograph to constrain the output toward a specific spatial structure. In this workshop, ControlNet is used within ComfyUI node-based pipelines to apply architectural references and depth maps as generation inputs. ### What does the Pixel Interfaces: AI Workflows in ComfyUI workshop cover? Day 1 introduces diffusion model principles, ComfyUI setup, and hands-on text-to-image and image-to-image workflows, followed by exploration of ControlNet conditioning, LoRA adaptations, and how parameters such as model choice, guidance scale, step count, and sampler affect output quality and computational cost. Day 2 shifts to interface building: participants select a workflow from Day 1 and package it into a browser-based generative tool using HTML, CSS, JavaScript, Python, and Gradio, culminating in a final critique of completed interfaces. ### What technical background do I need to join the ComfyUI generative design workshop? Participants should have basic familiarity with Python and some comfort with web technologies such as HTML, CSS, and JavaScript, as the second day involves building browser-based interfaces that expose ComfyUI workflows to other users. Prior experience with ComfyUI or diffusion models is not required, as Day 1 builds these from first principles. The workshop is suited to designers and architects with some programming background who want to engage with local AI generation as a substantive design tool rather than a hosted service. ### What will I produce at the end of the ComfyUI workshop at PAACADEMY? By the end of the two days, participants will have a working browser-based generative design interface built with Gradio, HTML, CSS, JavaScript, and Python, connected to a ComfyUI diffusion workflow that accepts one or more architectural inputs and returns meaningful generative output. The interface could be a facade variation tool, a material explorer, a comic layout generator, a site atmosphere tool, or any pipeline that takes architectural inputs and makes them accessible to someone else without requiring them to understand the underlying workflow.