1. Be Specific, Not Vague
Instead of "a beautiful sunset", write "a dramatic sunset over the Pacific Ocean with orange, pink and purple clouds reflecting on calm water, shot from a rocky cliff, golden hour photography."
2. Front-Load Important Elements
AI models give more weight to words at the beginning of the prompt. Put your most important subject and style first.
3. Use Art Style References
"In the style of Studio Ghibli", "Art Nouveau poster", "Pixar 3D animation" — specific style references produce more consistent results than generic terms.
4. Specify Lighting
Lighting transforms everything. Common keywords: golden hour, blue hour, rim lighting, backlighting, studio lighting, neon, volumetric, dramatic shadows.
5. Control Composition
Use terms like: extreme close-up, bird's eye view, low angle, symmetrical, centered, rule of thirds, negative space, full body shot.
6. Add Technical Details
Camera terms improve photorealistic images: 85mm lens, f/1.4, shallow depth of field, bokeh, film grain, 35mm film.
7. Layer Your Description
Build prompts in layers: subject → environment → lighting → style → mood → technical specs.
8. Use Mood Words
Ethereal, ominous, serene, chaotic, nostalgic, futuristic — mood words guide the overall feeling.
9. Avoid Negation in Positive Prompts
Don't say "no people" — instead use negative prompts or simply describe what you DO want.
10. Iterate and Refine
Your first prompt rarely produces the best result. Refine by adjusting one element at a time.
11-15: Advanced Techniques
- 11. Use weight syntax for emphasis in SD: (detail:1.3)
- 12. Combine unexpected styles for unique results
- 13. Reference specific color palettes by name
- 14. Use time period references for consistent aesthetics
- 15. Automate with tools like PromptFrom to learn from existing images