modular prompt blocks

[rich atlas]

What do you mean?

"""
If you want next level: I can split this into modular prompt blocks (so you can dial intensity of anatomy vs anime vs realism depending on the generator).
"""

[leaden harbor]

Hi @rich atlas. I'm kapa.ai, your friendly support bot. I will try to answer your question by searching through the knowledge base. :wave:

---

The knowledge sources don’t mention “modular prompt blocks” in exactly the way your quote uses it, but they do describe a very similar idea: breaking a prompt into clearly labeled sections or “blocks” that you can reuse and adjust.

In OpenAI’s docs, this shows up as:

• Labeled sections in a system prompt (for realtime models), e.g. separate blocks for “Role & Objective”, “Personality & Tone”, “Tools”, “Instructions / Rules”, etc.[Realtime prompt structure]
• Structured prompt templates for images and video, where you split content into sections like scene description, cinematography, actions, dialogue, constraints, etc.[Sora prompt structure][Image prompting fundamentals]

In that context, “modular prompt blocks” means:

• You design your prompt as separate, swappable chunks (blocks) instead of one big paragraph.
• Each block controls a specific aspect (e.g., style realism vs anime, level of anatomical detail, camera/lighting, etc.).
• You can then dial up or down a given aspect by editing or replacing just that block, while keeping the rest the same.

So applied to your quote, it would mean: having distinct prompt sections for “anatomy detail”, “anime style”, and “realism style”, and adjusting those sections independently depending on what you want the generator to emphasize—just like the docs suggest doing with labeled sections for role, tone, actions, cinematography, and constraints.

Note: you can ask me a follow up question by @ mentioning me again :speech_balloon:

@rich atlas Kapa.ai is still learning and improving, please let me know how I did by reacting below.

[rich atlas]

Example Modularized Prompt that ChatGPT shows me.

[BASE SCENE]
seated female figure on beach at golden hour, wrapped linen cloth, asymmetrical contrapposto, soft introspective pose

[ANATOMY INTENSITY: HIGH]
clear skeletal landmarks, pelvis tilt, ribcage counter-tilt, visible structure, academic accuracy priority

[FORCE + WEIGHT OVERLAY: MEDIUM]
visible force arrows, center of mass, weight distribution lines, but not overwhelming

[ANIME LINEWORK: MEDIUM]
rinkakusen contour lines, slight stylization, controlled kyojaku variation

[REALISM / MATERIAL: HIGH]
photoreal skin, subsurface scattering, fabric physics, sand compression, cinematic lighting

[ANNOTATIONS: LOW / CLEAN]
minimal handwritten notes, only key observations, not dense

[LIGHTING STYLE]
golden hour, warm rim light, cool shadows, soft falloff

[NEGATIVE]
no exaggeration, no glamour, no clutter, no distortion

@leaden harbor
Example Modularized Prompt that ChatGPT shows me.

[BASE SCENE]
seated female figure on beach at golden hour, wrapped linen cloth, asymmetrical contrapposto, soft introspective pose

[ANATOMY INTENSITY: HIGH]
clear skeletal landmarks, pelvis tilt, ribcage counter-tilt, visible structure, academic accuracy priority

[FORCE + WEIGHT OVERLAY: MEDIUM]
visible force arrows, center of mass, weight distribution lines, but not overwhelming

[ANIME LINEWORK: MEDIUM]
rinkakusen contour lines, slight stylization, controlled kyojaku variation

[REALISM / MATERIAL: HIGH]
photoreal skin, subsurface scattering, fabric physics, sand compression, cinematic lighting

[ANNOTATIONS: LOW / CLEAN]
minimal handwritten notes, only key observations, not dense

[LIGHTING STYLE]
golden hour, warm rim light, cool shadows, soft falloff

[NEGATIVE]
no exaggeration, no glamour, no clutter, no distortion

[leaden harbor]

Let me check the knowledge base for you again. One moment! :books:

---

The knowledge sources don’t mention this exact “modular blocks” style, but they do recommend the same underlying idea: splitting prompts into clear, labeled sections you can tweak independently.

For example:

• For ChatGPT / text prompts, OpenAI suggests headings like Context / Instructions / Constraints, each as its own block.[ChatGPT prompt guide]
• For Sora (video), they use blocks like Prose scene description, Cinematography, Actions, Dialogue.[Sora prompt structure]
• For images, they recommend a consistent order such as background/scene → subject → key details → constraints, and using short labeled segments for complex prompts.[Image prompting fundamentals]

Your example:

[BASE SCENE]
...

[ANATOMY INTENSITY: HIGH]
...

[FORCE + WEIGHT OVERLAY: MEDIUM]
...

[ANIME LINEWORK: MEDIUM]
...

[REALISM / MATERIAL: HIGH]
...

[ANNOTATIONS: LOW / CLEAN]
...

[LIGHTING STYLE]
...

[NEGATIVE]
...

is essentially applying that same pattern to drawing/image prompts:

• Each bracketed label ([BASE SCENE], [ANATOMY INTENSITY: HIGH], etc.) is a module that controls one dimension (pose, anatomy accuracy, stylization, realism, annotations, lighting, negatives).
• To “dial intensity,” you’d change the content or level in that block (e.g., ANATOMY INTENSITY: LOW with softer anatomical notes, or ANIME LINEWORK: HIGH with stronger stylization) while leaving the rest unchanged.

This matches OpenAI’s guidance to:

• Use clear sections so the model can “find and follow” them.[Realtime prompt structure]
• Keep a skimmable template for production prompts instead of one long paragraph.[Image prompting fundamentals]

So your modular prompt is a specialized, art-focused version of the structured prompt templates shown in the docs.

Note: you can ask me a follow up question by @ mentioning me again :speech_balloon: