rev0 / PortfolioYordan Vakarelov
PracticeBuilt ToolsLearningDesign Studies
All pages

Practice

About & CVStrategy & LeadershipPractice HeatmapTools & AgentsVisual IntelligenceTraining & AdoptionDecisions & Development

Built Tools

Door BuilderDesktop Tools

Learning

Workflow to TutorialLearning the Revit API

Design Studies

Basalt HouseThe Prompted InteriorDrawing with CodeStyle Studies

Design Studies / The Prompted Interior

The Prompted Interior

A professional guide to AI-assisted interior design. Five modular workflows turn ideas and references into concept briefs, photoreal interiors, specification-ready material mood boards and subtle motion studies.

vague to vivid
The Prompted Interior | From Idea to Image with ChatGPT & Midjourney
PublishedSold via Gumroadv01 · Oct 2025

The guide began as an independent study into AI-assisted interior mood boards. Testing tools and approaches turned it into five workflows that each do one job well and can be chained into a pipeline.

ChatGPT structures the brief and writes the design narrative; Midjourney explores, renders and builds the boards; Gemini’s image model handles the more controlled edits, such as mapping a palette onto a clay render.

5Modular workflows
2Custom Midjourney --p profiles
37Pages, plus full markdown prompts

Showcase

The guide in fifty seconds

The Prompted Interior | Guide showcase, with sound

How it fits together

Initial instruction to animated board

An initial instruction sets scope, tone and style rules for the session. From there, the workflows run as a set or on their own: a space generator writes the concept, a board narrative expands its materials, and the image, study and motion workflows take it into pictures.

Workflow map | The Prompted Interior, pp. 7-8

Workflow 01

Concept Curation

Generates or refines a room concept, then turns it into a design narrative and ready-to-visualise prompts. Three prompts run as a set or individually.

Input
Text instructions, short or extensive
Delivers
  • Initial instruction: intent, role and constraints
  • Space generator: full narrative, from materials and colour strategy to spatial logic, signature features and furniture, plus two hero-shot prompts
  • Board narrative & design: expanded materials, finishes and accents, with mood board prompts in two styles, Abstract Realism and Stable Simplicity
Saffron Washi Galley | Concept 2, Copenhagen townhouse kitchen
Hero shot | Space generator
Mood board | Board design
Reference | Single input image
Board | Built from the reference

Workflow 02

Image to Mood Board

Expands one reference image into a structured design study, reading both its physical composition and its sensory character, then produces like-for-like imagery and a matching material board.

Input
Text instructions and one or more reference images
Delivers
  • A full design narrative: palette, materials, lighting plan, signature feature and practicality
  • A like-for-like prompt that honours the reference composition and tone
  • Prompts for similar visuals and a material board that reflects brief and reference

Workflow 03

Image to Image

Maps the material and colour DNA of reference images onto a white-card or clay render, keeping the geometry and reinterpreting the finish.

Input
Text instructions, a clay or white-card base render and at least one reference
Delivers
  • A base-image description: openings, built-ins, lighting cues and constraints
  • A concept summary with palette, materials and atmosphere
  • Prompts with short and extended variants for reinterpretations and a matching board
Base | Clay render
Reinterpretation 01 | Walnut and teal
Reinterpretation 02 | Coastal blue
Reinterpretation 03 | Evening
Four-cube study | Linen, velvet, walnut, bouclé
Three-material study | Cork, eucalyptus, brass, rattan
Prompt · four-cube study

Flat lay, interior design mood board with 4 same size rough texture material cubes next to each other made out of dusty blue linen drapery, plush ochre yellow velvet, richly-grained walnut wood, textured bouclé upholstery, shot against a grey studio board background with Fujifilm X-T5 100mm, direct bright sunlight, deep shadows, contemplative stillness atmosphere --no low quality, bad quality, deformed geometry, unfinished, strange, weird, text, labels, watermarks, logos, branding, nudity, people, body parts --ar 1:1 --exp 2 --quality 2 --profile dxszxc4::2 vu1zsv8 --stylize 125

Workflow 04

Material Studies

Turns up to four material inputs into an editorial side-by-side board that tests cohesion, texture and colour balance in one clean composition.

Input
Text instructions
Delivers
  • A Midjourney prompt for 2, 3 or 4 side-by-side materials, with consistent scale, soft shadows and optional captions

Workflow 05

Animated Images

Adds tasteful, subtle motion to still interiors and mood boards, from a single start frame or a start and end pair. One prompt uses Gemini to build and animate a miniature model of the room.

Input
Text instructions and a start frame, or start and end frames
Delivers
  • Short interior animations that bring hero stills to life
  • Animated mood boards for reels, stories and presentation decks
Miniature model | Built and animated with Gemini
Mood board disassembly | Start and end frames
Sun study | Light moving across a board
Board to image 01 | Materials become the room
Board to image 02 | Materials become the room
Disassembly | Board pulled apart

Why it works

Controlled creativity

01

Vague to vivid

Rough ideas become concrete briefs and visuals quickly, with the reasoning written down.

02

Repeatable

Tuned settings, two custom profiles and a seed library refined for material boards keep quality consistent.

03

Modular

Each workflow runs on its own or as part of a pipeline, so it fits the project rather than the other way round.

04

Presentation-ready

Editorial boards and subtle animations come out ready for decks, reels and client packs. Defaults work; advanced options unlock more control.

From the guide

Selected images

About the guide

Personal work, independent of practice. Optimised for ChatGPT-5 Thinking and Midjourney v7.1, with some use of Gemini’s Nano Banana image model. Outcomes vary by nature; the prompts are springboards for iteration rather than fixed blueprints.

Keep exploringTraining & Adoption ↗Style Studies ↗
Yordan Vakarelov · Design Studies · 13 / 15
← PreviousBasalt HousePortfolio ↑Front pageNext →Drawing with Code
0