Stable Diffusion Prompt Reader

A good AI prompt provides clear, detailed instructions to the AI model on what you want it to generate. The prompt should communicate the desired content, style, composition, and other qualities of the output.

Here are some tips for writing effective AI prompts:

  • Be specific. Include concrete details about what you want the AI to generate. Vague prompts lead to vague results.
  • Give examples. Provide sample images, texts, or other references to illustrate what you have in mind. This gives the AI more context.
  • Use descriptive language. Adjectives, adverbs, and descriptive phrases will produce richer and more accurate outputs.
  • Specify a style. Tell the AI what artistic style, genre, time period, or aesthetic you want.
  • Set parameters. Define the medium, size, orientation, framing, aspect ratio, etc. This controls the composition.
  • List themes. Identify topics, concepts, emotions, or other themes to inform the AI’s creation.
  • Be concise. Shorter prompts tend to work better than long, complex ones. Remove unnecessary words.

Prompt Structure

Well-structured prompts make it easier for the AI to understand what you want. Here is an effective prompt formula:

Subject and medium > Style/genre > Themes/mood > Setting/era > Composition > Quality

For example: “An oil painting of a rainy city street scene with a nostalgic, romantic atmosphere; narrow depth of field; textured brush strokes; cinematic lighting; highly detailed; IMAX film aspect ratio.”

Breaking down the components:

  • Subject and medium: An oil painting
  • Style/genre: Rainy city street scene
  • Themes/mood: Nostalgic, romantic atmosphere
  • Setting/era: No specifics given
  • Composition: Narrow depth of field
  • Quality: Textured brush strokes, cinematic lighting, highly detailed, IMAX film aspect ratio

Prompt Engineering Strategies

You can further refine your prompts using these advanced techniques:

  • Weighted modifiers – Stress important attributes by repeating them or adding “highly” or “extremely”
  • Style transfer – Combine multiple styles in one prompt
  • Imaginary realism – Mix fantastical elements with photographic realism
  • Textual inversion – Insert an image and ask AI to make something “related to” or “inspired by” it
  • Nested prompts – Break long prompts into shorter sentences
  • Memory cues – Remind the AI of important elements from previous prompts

Stable Diffusion Prompt Examples

Here are some prompts engineered specifically for Stable Diffusion:

Photorealistic Portraits

Extremely detailed digital painting of a beautiful young woman with long blonde hair, bright green eyes, dark eyeliner, red lips, wearing a black leather jacket; cinematic lighting; depth of field; 8k resolution

Breakdown:

  • Subject and medium: Extremely detailed digital painting
  • Style/genre: Photorealistic portrait
  • Themes/mood: Beautiful, cinematic
  • Setting/era: Contemporary
  • Composition: Shallow depth of field
  • Quality: 8k resolution

Futuristic Cityscapes

A futuristic city of towering skyscrapers and flying vehicles, vertical urban farm on side of building covered in lush greenery, utopian metropolis, volumetric lighting, cinematic perspective, Unreal Engine, trending on Artstation HQ

Breakdown:

  • Subject and medium: A futuristic city
  • Style/genre: Skyscrapers, flying vehicles, vertical urban farm
  • Themes/mood: Utopian metropolis
  • Setting/era: Futuristic
  • Composition: Cinematic perspective
  • Quality: Volumetric lighting, Unreal Engine, trending on Artstation HQ

Useful Prompt Engineering Resources

I hope these AI prompt examples and resources help you create amazing AI artworks! Let me know if you have any other questions.