Getting good results
Last updated
The two things you control that move output quality the most: the photo you upload and the instruction you write.
Most quality problems are input problems
When a render disappoints, the instinct is to rewrite the prompt. It is usually the photograph. The model reads the room's geometry, light direction, and surfaces out of the image you gave it, so a dark, tightly cropped, or heavily angled shot removes information no wording can put back. A wide, well-lit photo taken with your back to the longest wall, with both the floor and the ceiling line in frame, prevents more problems than any phrasing will fix. That is why photographing a room comes before writing a good prompt in this section, and not the other way round.
When to stop tuning and regenerate
Generation is not deterministic, and that trips people up more than any other property of the tool. The same photo and the same instruction produce a different image on every run, which means a mediocre result is often just a mediocre roll rather than evidence your prompt is wrong. The habit that follows is to generate two or three and pick, rather than rewriting after every attempt. Why two runs give different results covers this properly, including how to tell ordinary variation apart from a result that is consistently wrong. When a render fails outright, or comes back with a warped floor or a smeared edge, that is a different problem and has its own page.
How should I take the photo?
The upload is the single biggest lever you have on output quality, and it is worth two minutes of care. The model can only redesign what it can see, so a photo that reads clearly to your eye also reads clearly to the model. None of this needs a real camera. A recent phone is fine.
What should I write in the prompt box?
The prompt is optional. Choosing a style and a mode is often enough, and a bad prompt does more damage than no prompt at all. Where a prompt earns its place is in the specifics a style name cannot carry: the sofa you want to keep, the oak floor that should survive, the exact green you have already bought paint for.
Why does the same photo and prompt give a different image each time?
Because generation samples a new result each time rather than looking one up. Run the same photo and the same words twice and you get two different rooms, both valid. This is the single most useful thing to understand about working with the product, and it is why the habit that produces good output is generating two or three and picking rather than rewriting after every attempt.
My design failed or came back wrong. What now?
A design that fails does not cost you anything. The credit is returned automatically when the generation is marked failed, without you having to ask. A design that succeeds but disappoints is the more common case, and it is nearly always traceable to the input rather than to bad luck.
How do I design several rooms in one property?
One room at a time. Every mode takes a single photograph and returns a single photograph, so a five room property is five separate jobs rather than one batch. Nothing carries over between them automatically, which means consistency across rooms is something you create through how you write the instructions rather than something the product maintains for you.
Other sections