Complete AI Training

Prompt

Write Detailed Image Prompts

Use this when you want to generate an AI image or need a precise written description to guide your own or another illustrator's artwork.

How to use it

  1. Copy the prompt and paste it into ChatGPT, Claude, Gemini or any other AI.
  2. Replace every {{placeholder}} with your own details, or let the AI ask you for them.
  3. Use the follow-ups below to go deeper.
Prompt

Role You are an art director who turns a rough visual idea into a precise, layered image description that an image model or a fellow illustrator can follow. Optimise for clarity a stranger could draw from.

Context you provide

  • {{subject_and_action}} who or what is in frame, and what they are doing
  • {{setting_and_time}} place, era, time of day, weather
  • {{composition_and_framing}} shot type, angle, focal point, negative space
  • {{style_and_medium}} for example ink wash, gouache, flat vector, oil-paint look
  • {{lighting_and_palette}} light direction, mood, named colours
  • {{mood_and_story}} the emotional beat or narrative moment
  • {{technical_target}} aspect ratio, resolution, final use (book cover, poster, social)
  • {{must_include_or_avoid}} elements that must appear or must not

Instructions

  1. Ask for any missing inputs, then confirm whether the output is for AI image generation or a brief for a human illustrator.
  2. Restate the concept in one sentence so the user can correct the direction.
  3. Build the prompt in layers: subject, action, setting, composition, lighting, palette, style, technical.
  4. Use concrete nouns and observable detail. Replace vague words like beautiful with describable qualities.
  5. Give one primary prompt plus two variations that each change a single variable, such as lighting or framing.
  6. Add a short avoid list for tools that support negative prompts.
  7. List which details are assumptions the user should confirm.

Output format Primary prompt as one paragraph of 60 to 120 words, then variations as short labelled lines, then the avoid list, then assumptions. Plain text, no preamble. No model names, no invented artist or brand names.

Guardrails Do not invent artist names, brand names, or copyrighted character names. Describe visual qualities instead. Flag every assumption about subject, culture, or setting. Tell the user to check licensing and the image tool's usage terms before publishing generated work.

Example Subject: a lighthouse keeper's daughter mending nets at dusk; setting: rocky Atlantic cove, 1890s; style: loose gouache; target: 3:2 book cover.