Prompting for Image generation

The prompt box lets you describe your design intent in plain language, giving Motif's AI more context to produce accurate, targeted results. You'll write prompts in two situations:

• To guide a render — when you render from a view, photo, or sketch, your prompt adds context on top of what Motif already detects in the input.

• To generate from text — when you have no starting image, your prompt describes the entire image. See → 📄 Generate an image from a text prompt.

The principles below apply to both. The main difference is how much work the prompt is doing: with a starting image it supplements what's already there; with text only it has to carry the whole description — subject, materials, setting, and viewpoint.

Using a prompt to guide a render

Motif will automatically detect colors and materials defined in your model and incorporate them into the render. However, because AI works through interpretation rather than exact replication, results may differ from your specifications. Use the prompt box to:

• Override or specify materials and colors

• Describe the building type or program (especially for early-stage or conceptual models)

• Add entourage like people, vegetation, or furniture

• Clarify the camera angle or viewpoint

• Guide the style or atmosphere beyond what modifiers offer

Start by simply clicking on the image you want to render from and click the prompt button to start typing your prompt:

Result from the above example:

Example prompts

Here are simple, effective prompts to get you started:

• Soccer stadium clad in aluminum

• Apartment building in mass timber, lush green roof

• Render this apartment building in Miami deco colors

• Fully clad in small tiles

• Chairs in random pattern of dark green and light green — floor, walls and ceiling in dark grey

• Lush roof garden of tall residential tower

Clarifying the camera view

If the AI doesn't seem to understand the camera angle — common when there is little visual information in the input — add directional cues to the prompt:

• aerial view

• aerial view, sky is not visible

• ground is not visible

• detail shot of top of building, ground is not visible

Next steps

• 📄 Generate an AI render from an image or view

• 📄 Edit rendered images

• 📄 AI Rendering tips, best practices, & known issues