Back to library

🎨Master Text-to-Image Prompt Craft

Build an internal recipe for prompting diffusion models — subject, medium, style, lighting, weight, negative — so you can generate brand-aligned images on demand instead of copying random prompts from marketplaces.

Foundations14 drops~2-week path · 5–8 min/daycreativetechnology

Phase 1How Diffusion Models Actually Read Your Prompt

Discover why specificity steers diffusion models

4 drops
  1. A cat is not a prompt — it's a coin flip

    6 min

    A cat is not a prompt — it's a coin flip

  2. Every prompt is four ingredients in a trench coat

    6 min

    Every prompt is four ingredients in a trench coat

  3. Photographers don't say 'nice light' — neither should you

    7 min

    Photographers don't say 'nice light' — neither should you

  4. Copying a stranger's prompt borrows their taste, not their skill

    6 min

    Copying a stranger's prompt borrows their taste, not their skill

Phase 2One Variable at a Time

Iterate one variable at a time per drop

5 drops
  1. Pick a base prompt boring enough to improve

    6 min

    Pick a base prompt boring enough to improve

  2. Medium changes everything — try four to feel it

    7 min

    Medium changes everything — try four to feel it

  3. A 24mm wide-angle is a different story than an 85mm portrait

    7 min

    A 24mm wide-angle is a different story than an 85mm portrait

  4. Borrow a style without naming a living artist

    7 min

    Borrow a style without naming a living artist

  5. Weights and negatives are the steering wheel's fine adjustments

    7 min

    Weights and negatives are the steering wheel's fine adjustments

Phase 3Match the Prompt to the Model

Match prompt style to each model's strengths

4 drops
  1. Midjourney rewards vibe — but only if you respect its house style

    7 min

    Midjourney rewards vibe — but only if you respect its house style

  2. Stable Diffusion rewards precision — and punishes laziness

    8 min

    Stable Diffusion rewards precision — and punishes laziness

  3. Flux speaks natural language — write the caption, not the prompt

    7 min

    Flux speaks natural language — write the caption, not the prompt

  4. Pick the model before you write the prompt, not after

    7 min

    Pick the model before you write the prompt, not after

Phase 4Ship a Brand Hero Image

Generate and document a brand hero image

1 drop
  1. Generate a brand hero image and document every revision

    8 min

    Generate a brand hero image and document every revision

Frequently asked questions

What does subject-medium-mood-lighting actually mean in a prompt?
This is covered in the “Master Text-to-Image Prompt Craft” learning path. Start with daily 5-minute micro-lessons that build from fundamentals to hands-on application.
Why do four-word prompts produce generic AI images?
This is covered in the “Master Text-to-Image Prompt Craft” learning path. Start with daily 5-minute micro-lessons that build from fundamentals to hands-on application.
How is Midjourney different from Stable Diffusion for prompting?
This is covered in the “Master Text-to-Image Prompt Craft” learning path. Start with daily 5-minute micro-lessons that build from fundamentals to hands-on application.
What is a negative prompt and when should I use one?
This is covered in the “Master Text-to-Image Prompt Craft” learning path. Start with daily 5-minute micro-lessons that build from fundamentals to hands-on application.
How do prompt weights work in diffusion models?
This is covered in the “Master Text-to-Image Prompt Craft” learning path. Start with daily 5-minute micro-lessons that build from fundamentals to hands-on application.