Archshaper vs Shapezo: A Prompt to Place Workflow for Early Architecture

I think about early architectural visualization as two separate inputs: a design brief and a geographic boundary. Archshaper works from the design brief. Shapezo works from the selected map area. Comparing them this way makes it easier to choose a starting point and to keep the generated output in the right role.

1. Write a Small Design Brief Before opening Archshaper, I collect four things: building type, site conditions, primary use, and target character. For example, I might describe a small fire station on a corner parcel, with public access from the main street, service doors on the side, and a durable brick exterior. The purpose is to give the AI a clear visual direction, not to replace a drawing set. Archshaper can then help me explore exterior form, interior atmosphere, materials, and style. I keep the first prompt focused. If I include every preference at once, it becomes difficult to see which part influenced the result.

1. Set a Geographic Boundary When Context Is the Unknown Shapezo uses a map selection as its starting point. I zoom to the area, draw a rectangle around the part I want to study, and let the AI generate a model for that region. I include enough surrounding streets and buildings to make the site legible, but avoid selecting a region so broad that the project becomes a small detail. This workflow is useful when I need a quick model of place. It gives me an overview to inspect before I decide which site relationships deserve a closer look. I still treat the output as generated context, not as a verified survey or parcel record.
2. Separate Visual Variables When I use Archshaper, I change one or two variables at a time. I might keep the building form stable while comparing timber cladding with pale masonry. Then I can change the entry or roof shape in a separate pass. This makes the review more useful because I can explain what changed between images. When I use Shapezo, I document the selected map extent. If I change the box between generations, I note it. Otherwise, a difference in the model could come from a different area rather than from a design choice.

1. Convert Reactions Into Checks Generated images are good at prompting reactions. “The entrance looks hidden” is a useful observation. I turn it into a check: where is the door, what path leads to it, and can it be seen from the street? “The room feels cramped” becomes a question about dimensions, furniture clearances, and circulation. The image helps me form the question. It does not answer the technical version by itself. For structure, code, construction, accessibility, or exact site fit, I bring in the appropriate plans, data, and professional review.

1. Choose the Handoff by Artifact If I need to discuss a facade direction, room mood, or concept option, an Archshaper visual may be a useful communication artifact. If I need a broad 3D understanding of a mapped area, the Shapezo model may be more relevant. In either case, I label the output as an early visual study and keep the brief or map boundary attached to it. I also keep the next tool in mind. A concept image, a map generated model, a measured CAD model, and a construction document are different deliverables. Mixing them up leads to vague reviews and incorrect assumptions about what has been verified. My Working Rule I use Archshaper when the main input is an architectural intention. I use Shapezo when the main input is a place selected on a map. Then I use the image or model to sharpen the next question, not to skip it. That gives me a fast first pass while preserving the checks that real design decisions need. A Simple Review Record For a team review, I keep the original brief, the chosen image, and one sentence about the decision still open. That makes feedback more concrete. Instead of asking whether a picture feels finished, I can ask whether the entrance reads from the sidewalk or whether the selected context includes enough of the surrounding block.