Interior AI revenue: Pieter Levels' room redesign tool

Repainting a living room, buying a new sofa or staging an apartment for sale all share one problem: you have to spend the money before you can see the result. Interior AI flips that order. Upload one photo of a room, pick a style, and seconds later you see the same room redesigned. It was built by Pieter Levels in 2022, in the first weeks after open image models became public, and it is one of the clearest examples of a product whose marketing is simply its output.

Why did Pieter Levels build Interior AI?

The technical trigger was image-to-image diffusion. These models made it possible to keep the shape of a room, its walls, windows and layout, while changing everything else: furniture, colours, materials and lighting. That is exactly what people deciding on a redesign need.

Levels saw an obvious paying audience: homeowners, renters, real-estate agents and interior designers who want to preview a style before spending money. The pain is uncertainty. A designer’s mood board or a 3D render costs time and money; a quick AI preview costs almost nothing and answers the question “would this look good in my room?”

How does Interior AI work?

The flow is upload-and-pick-a-style. A user photographs a room, uploads it, and chooses a style. A diffusion model, run through a hosted GPU API, generates redesigned versions that preserve the room’s structure. The user can try several styles and compare.

That is the entire product, and its narrowness is a strength. There is no account setup ritual, no floor-plan editor and no learning curve. The value is visible within one try, which also makes it easy for users to share the before-and-after results.

How was Interior AI built?

Levels used the same lightweight approach as his other products: a PHP and jQuery front end on a single server, with diffusion models run through a hosted GPU API. He does not train or host his own GPU fleet, which keeps the operation small enough for one person.

He launched a simple upload-and-pick-a-style page in the autumn of 2022, shortly after Stable Diffusion became public, and shared results on X. The public launch date is slightly uncertain: directory listings disagree between September and October 2022. Team size is not disclosed separately; Levels runs his products largely alone.

How did Interior AI market and price itself?

Before-and-after images are the marketing. Levels posted them on X, and the product was listed on Product Hunt and many AI tool directories. Product Hunt launch figures for Interior AI are not disclosed in the sources this case relies on (Hunted Space, Product Hunt).

Pricing is paid plans for a set number of renders per month. That structure fits the likely customer mix: a homeowner might render a handful of rooms once, while professional users such as agents and stagers, who are the natural heavy users, need many renders every month.

What challenges does Interior AI face?

The biggest challenge for anyone studying Interior AI is the lack of reliable numbers. Monthly revenue is not disclosed, total revenue is not disclosed, and team size is not disclosed. Figures in circulation come from unverified secondary blogs, so they are not used here.

There is also a product challenge common to many early diffusion tools: the underlying models are available to anyone. The technical moat is thin, so distribution and focus on a clear vertical carry more weight than the model itself. Interior AI also runs alongside Photo AI as part of Levels’s portfolio from 2023 onward, so it competes for the same founder’s attention.

How much revenue does Interior AI make?

Interior AI’s revenue is not disclosed. No primary source, such as a founder post with a date or a payment-processor screenshot, was found for a monthly or total figure, so this case reports “Not disclosed” rather than repeat estimates from secondary blogs.

What is documented is the timeline. The product launched publicly in September 2022 (with some listings saying October 2022), and from 2023 onward it has been run alongside Photo AI, the larger product in the same portfolio, which does have a founder-reported revenue milestone.

Why did Interior AI work?

The product picked a vertical where one picture shows the value instantly. A before-and-after pair of images communicates the whole offer without a single word of explanation, which makes every result a potential post, and every post a potential customer.

It also targeted buyers who pay for speed. Real-estate and design professionals already spend money on staging, renders and presentations; a faster, cheaper preview fits an existing budget line. Compare HeadshotPro, which applied the same “one image proves it” logic to business headshots.

What can founders learn from Interior AI?

Pick a vertical where one picture shows the value instantly. If your result can be understood in a single glance, your users will market it for you.

Real-estate and design professionals pay for speed. Look for professionals who already pay for a slow, expensive version of what AI can now do quickly.

And treat headline revenue figures with caution when there is no primary source. Interior AI is widely discussed, but without a founder-reported or verified figure, the honest answer to “how much does it make?” is “not disclosed”.