Archmaster
AI Design Chat

Redesign Your Room by Describing It

Tell the assistant what you want. It looks at your photo first, works out the room, the materials and the light, then writes the prompt for you.

The AI design chat listing the furniture it found in a studio render

What Is AI Design Chat?

AI Design Chat is a second way to use Archmaster: instead of picking a feature and a style from a form, you upload a photo and describe what you want in your own words.

How designing by chat works, step by step

The assistant looks at your photo before it writes anything, so the prompt it builds names your actual cabinetry and your actual light.

01

Upload a photo

Drop in any room, facade or garden photo. JPG, PNG or WebP up to 12 MB. This is the only thing the assistant needs before it can start.

  • JPG, PNG, or WebP up to 12 MB
  • The only thing the assistant needs before it starts
02

Say what you want

Plain language is enough. The assistant reads the photo, tells you what it found, and turns your sentence into a prompt that names the real materials and lighting in your space.

  • Reads the room type, materials, and light before writing a prompt
  • Names your actual cabinetry and windows, not a generic label
03

Keep going from there

Ask for a different style, another angle, a video, or the products in the render. Every follow-up keeps the context of the photo and everything already said about it, so you never re-explain the room and never paste the same description twice.

  • Ask for a different style, another angle, or a video
  • Every follow-up keeps the context of the conversation

The problem

The prompt is the part most people get wrong

Image models reward prompts that name materials, light, and framing in the language the model was trained on. Most people write a sentence about what they want, get something generic back, and conclude the tool does not work. The gap is not taste, it is vocabulary.

  • A short prompt returns a generic room, not your room
  • Learning what a model responds to is a skill nobody asked to acquire
  • Form fields cover the common cases and nothing outside them
  • Every re-roll costs a credit, so guessing is expensive

What this does

Say it plainly, let the assistant write the prompt

Upload a photo and describe what you want in ordinary words. The assistant reads the photo first, tells you what it found, and turns your sentence into a prompt that names the real materials and lighting in your space, then runs it.

  • Reads the room type, materials, and light from the photo before writing anything
  • Names your actual cabinetry and windows instead of a generic label
  • Follow-ups keep the context, so you never re-describe the room
  • 1 credit per message, plus the cost of what it generates
A render with a tappable dot open on a navy bath towel product card
Every piece becomes tappable
Google Lens shopping results showing navy towels from IKEA with prices
Real sellers, real prices

100,000+

Rooms redesigned

5,000+

Happy users

No data storage

Photos are processed and deleted β€” never kept or sold.

Secure payments

Checkout handled entirely by Stripe.

What people do with it

I tried four styles on my living room before picking one. Seeing the actual result before spending a dollar on furniture saved me from a mistake.

Sarah Chen

Homeowner

Staged three vacant listings in one afternoon. What used to take a week and a rental invoice now takes minutes per room.

Marcus Webb

Real estate agent

I use it to show clients three directions in the first meeting instead of describing them. They sign off faster because they're looking at their actual room.

Priya Patel

Interior designer

Turned a rough kitchen sketch into something I could actually show a contractor. Cut a week off getting quotes.

Tom Reyes

Homeowner

Tested a paint color and new furniture layout before asking my landlord about anything. No surprises, no wasted money.

Elena Novak

Renter

FAQ

How is this different from the normal workspace?

The classic workspace sends your text straight to the image model with a fixed template. Chat analyses the photo first and writes the prompt around what is actually in the room, which is usually the difference between a render of your space and a render of a generic one.

Do I have to re-upload my photo for every request?

No. The conversation remembers the photo it is working on, so a follow-up like 'now try it Mediterranean' needs nothing but the sentence.

What does it cost?

Each message costs 1 credit. Any render the assistant makes costs the same as in the classic workspace: 1 credit for a redesign, 3 for an alternate angle, 5 for a video.

Can it do everything the classic workspace does?

Yes. Interior and exterior redesign, sketch to image, upscaling, targeted edits, alternate angles, video, virtual staging, room planning, landscape design and floor plans are all available through the conversation.

Can I go back to the classic workspace?

Any time. The two share the same account, the same credits and the same render library, and switching is one click in either direction.

Stop filling in forms. Just say it.

Upload one photo, describe the room you want, and let the assistant do the part that decides whether the render looks like your home or somebody else's.

Start designing