Gen-AI-tecture: a bespoke, locally run ComfyUI generative-AI image creation and editing workflow for architectural education¶

Design · All designs
Evidence · 1 claim (1 for) · 1 study (1 review), q2 · 0 of 1 report an effect size · 1 claim rests on one study

Description¶

A bespoke generative-AI image workflow implemented as a localised ComfyUI environment in which students upload or select images, describe changes in natural language, and receive multiple visual alternatives. "The workﬂow is powered by the Flux 1 Kontext (dev) diffusion model (Black Forest Labs, 2025), which has been ﬁne-tuned with a curated set of interior-design images" using LoRA adaptation on the InteriorNet dataset. It offers three modes (text-driven, reference-image-driven, hybrid) and edits via in-painting that preserves surrounding context, acting as a visual co-pilot rather than replacing design thinking.

Design Implications¶

Context¶

Requirements¶

Runs locally within a ComfyUI node-based environment with the fine-tuned Flux 1 Kontext (dev) model
Used within scaffolded studio sessions with facilitator demonstration and structured design tasks

Constraints¶

Fine-tuned for residential interior scenarios (bedroom and living/dining room scenes), so outputs support early design discussions "without claiming the status of resolved architectural drawings" and may not generalise to other design typologies or less supported studio environments

Target Learners¶

Undergraduate architecture students (Levels 3-5)

Learning Goals¶

Architectural design ideation and visualisation
AI-handling skills for professional practice

Claims¶

Multi-step AI operations, not AI use alone, are positively associated with creativity of final design work [+W]

Examples¶

-

Key Sources¶

Kapsalis, T. (2026). Gen-AI-tecture: using generative AI to support architectural students in design tasks. https://orcid.org/0000-0003-1598-6426

Page metadata

Field	Value
`type`	`design`
`id`	`gen-aitecture-comfyui-inpainting-workflow`
`title`	`Gen-AI-tecture: a bespoke, locally run ComfyUI generative-AI image creation and editing workflow for architectural education`
`description`	`A bespoke generative-AI image workflow implemented as a localised ComfyUI environment in which students upload or select images, describe changes in natural language, and receive multiple visual alternatives.`
`status`	`review`
`generated`	by `process:settle-candidates` on `2026-10-10`
`trust`	unverified — no one has confirmed this page's content
`sources`	1 — `kapsalis-2026`

Frontmatter is the page's machine-readable record — see the Schema & Guide.