Skip to content

Image Studio overview

What will you draw today? Nine genres of architectural imagery, made from a sentence or from your model.

ProGuide · 2 min read

An architect draws a dozen kinds of picture before anything is built: the exploded axo that explains a structure, the watercolour a client frames, the planting section that settles an argument, the white-card model that makes a crit go quiet. The Image Studio is built around those pictures, not around AI models. You say what you're making, how it should look and what it is, and the studio does the rest.

6 stylesDiagramsDraw the idea, not the render: axos, sections and step-by-step diagrams from a sentence or a viewport capture.
8 stylesSketch & Hand MediaYour model, redrawn the way architects think: in pencil, ink and wash.
7 stylesInteriorsWalk into the room before it exists: try its light, its materials and the life it will hold.
8 stylesLandscape & EcologySee a landscape through its seasons, its weather and its daily life before a single plant goes in the ground.
8 stylesExterior VisualisationPut the building on its street before anyone pours a foundation.
7 stylesSite PhotomontageStart from the street as it is, and change only the thing you're proposing.
8 stylesModel PhotographyThe crit-table model you didn't have the week to build.
7 stylesFacade & MaterialTry the skin before you choose it: the same module, different materials, twenty-five years on.
8 stylesConstruction & PhasingShow how it goes up: the frame, the crane, the first panel on the hook.

How it works#

You pick a genre and a style and describe the subject. The platform composes the prompt, picks the model for that style and draws the image. One image comes back, labeled as imagined.
  1. Pick a genre. Diagrams, sketches, interiors, landscapes, exteriors, photomontage, model photography, facades or construction. Each opens with its own styles and choices.
  2. Pick a style and a few choices. An ink-line axonometric or a soft pastel plan; morning light or dusk; four panels or one. The studio remembers them per genre.
  3. Say what it is. A sentence about the subject: "a four-storey timber school around a courtyard; the roof lifts to bring light into the hall". Leave the look to the style.
  4. Start from your model, if you like. Capture the Rhino view, or attach a sketch or a site photo. See Starting from your model.
  5. Generate. The image appears on the canvas. The Prompt button under it shows exactly what the model was asked, word for word.

Open the studio from the Wrengler panel: Systems, then Generate image in the header. Or ask the assistant with the Image chip, and it proposes a genre, a style and a subject for you to generate. In a project's chat on the web, Generate makes the image right there from your description (there is no Rhino view to start from), and Save to project keeps a labeled copy in the project's files with a note of how it was made beside it.

What goes into a prompt#

Only two things: the wording of the genre, style and choices you picked, and the description you typed. The studio never adds study results, project documents, project facts or file names. A sunny courtyard in an image says nothing about whether the courtyard gets sun; for that, run the Sunlight study.

What you get#

  • One image per request, on the canvas, with the images you made earlier in the session beside it.
  • The prompt and the caveats under every image.
  • An export with a label burned into the pixels that says it is AI-generated and not a simulation result, the same in its metadata for software to read, and a provenance file beside it. See Labels and provenance.

Note

Pro includes 100 images a month, with no key to set up and nothing else to pay. Every image counts as one, whichever style made it.