No commands, don’t use conversational language. Describe what you want to see.
❌ “change colour”, “remove the curtains”, “don’t make it dark”, ”make it bigger”
✅ ”a large red velvet sofa”, “bright and airy atmosphere”
Be clear and specific rather than using generic descriptions. Clearly describe the key elements how you want to see them.
❌ “a vase on a table”
✅ ”a rounded wooden coffee table with a blue glass vase on it”
The more precise you are, the more accurate the result.
❌ “a castle”
✅ ”a majestic castle on top of a mountain overlooking a winding river”
Order matters! AI prioritises the earlier terms in your prompt. Words placed at the beginning carry more weight, meaning that starting with the main subject ensures the AI understands the most important element first.
<aside> 🔥
"A neon-lit city street with glowing billboards and people walking. The futuristic skyline in the background features a towering glass skyscraper reflecting the sunset. The building’s sleek, modern design includes sharp angles and mirrored surfaces. Cinematic lighting, ultra-detailed, photorealistic, 8K architecture photography."
.png)
</aside>
<aside> 🔥
“A futuristic glass skyscraper towering over a vibrant city skyline, reflecting the warm hues of the sunset. The building’s sleek, modern design features sharp angles and mirrored surfaces. Neon-lit streets below are filled with people walking under glowing billboards. Cinematic lighting, ultra-detailed, photorealistic, 8K architecture photography.”
.png)
</aside>
The ideal prompt should include all essential elements you want to control. Any details not specified will be left to the AI’s interpretation and creativity.
<aside> 💡
Recommended formula:
[Main subject], [Context & Key features], [Atmosphere & Lighting], [Style], [Extra Keywords]
<aside>
“A modern villa with a swimming pool in front, surrounded by pine trees, natural day light, contemporary style, architectural photography, high resolution”
_(4).png)
</aside>
</aside>
Be descriptive but avoid complexity.
<aside> 💡
There is no strict rule that "shorter is better" or "longer prompts work best"—it all depends on clarity and structure.
</aside>
❌ “A futuristic high-rise tower standing at the heart of a bustling metropolis, its sleek glass facade reflecting the warm glow of the sunset. The skyline is dotted with other skyscrapers, while tree-lined boulevards and pedestrian-friendly spaces weave through the urban fabric. In the foreground, an expansive plaza with contemporary sculptures invites visitors to gather and admire the architectural marvel. This ultra-detailed image captures the vibrant energy of the city with cinematic precision, blending modern design with dynamic urban life.”
What’s ❌**?**
Phrases like "bustling metropolis," "contemporary sculptures," and "dynamic urban life" added unnecessary conceptual descriptions rather than guiding the AI on specific visual elements. The excess detail resulted in blurry and less defined architectural features.
✅ ”A futuristic glass skyscraper towering over a vibrant city skyline, reflecting the warm hues of the sunset. Surrounding urban landscape with tree-lined streets and a lively plaza, cinematic lighting, architectural visualisation, ultra-detailed”
What’s ✅**?**
Included all critical details without overloading it with unnecessary descriptions, retained all essential architectural details while improving clarity and processing efficiency.
In this example, the shorter version preserves all key architectural details while improving AI processing efficiency, ensuring that even small elements remain clear and well-defined.
Ensure that each element is clearly separated and precisely defined when describing patterns or materials. Avoid using multiple or too complex adjectives that may cause confusion. Clearly separate and define each object to ensure accurate representation.
_(7).png)