# [[02-diffusion_basics|Prev]] | Article 03 — Prompting fundamentals | [[04-image_as_source|Next]] Also known as [[diffusion_models#^overview-conditioning|conditioning]], writing a prompt is about giving the model clear instructions. Breaking down your request into specific components, gives more consistent results. ## The Anatomy of a Prompt ![[prompt_components#^overview]] [[prompt_components#The 6 Core Components|6 core components]] can be used as guideline to help structuring the prompt: [[prompt_components#^subject-core-component|Subject]], [[prompt_components#^medium-core-component|Medium]], [[prompt_components#^style-core-component|Style]], [[prompt_components#^composition-core-component|Composition]], [[prompt_components#^lighting-core-component|Lighting]] and [[prompt_components#^quality-core-component|Quality]]. ![[prompt_components#^anatomy-components]] ## The Power of Negative Prompts ![[negative_prompts#^overview]] ![[negative_prompts#Common Pitfalls]] ## **Advanced Control**: Weighting & Scheduling Once the basics are solid, you can introduce finer controls to the generation. These techniques let you balance the importance of concepts and words in the prompt and when they affect the generation. [[prompt_weighting|Weighting]] lets you amplify or soften specific concepts without removing them entirely. [[prompt_scheduling|Scheduling]] lets you change your instructions mid-generation, using [[diffusion_models#Understanding Steps|early steps control composition while late steps control detail]].