Open main menu
Blog HomeGuidesCreator Stories
Create AI Cartoons

Skin Tone Names for AI Prompts: Complete Guide (2026)

Master inclusive skin tone naming for AI prompts in 2026. Learn accurate descriptors, undertones, Fitzpatrick-inspired ranges, and bias-free examples

Table of Contents
Skin Tone Names for AI Prompts: Complete Guide (2026)
Do not index
Do not index
notion image
Getting skin tone right in AI-generated characters? It's one of the most overlooked skills in character creation. And it's costing you hours of wasted iterations, inconsistent results, and characters that drift from page to page.
This guide gives you a practical naming system, a copy-paste library of prompts that actually work, and a workflow for keeping skin tone locked down across dozens of scenes. Whether you're illustrating a children's book, building characters for educational content, or creating a social media series, you'll walk away with prompts you can start using today.

Why AI Models Struggle with Skin Tone Prompts

Most tutorials won't tell you this: AI image models don't store skin tone as a neat attribute like skin_tone = medium_brown. They predict pixels. And those pixels get affected by three things that constantly work against you:
notion image
Lighting and white balance. Warm sunset light pushes skin orange. Cool moonlight makes it look grey or blue. If your prompt doesn't separate "base skin tone" from "scene lighting," the model will actually change the character's skin color to match the mood of the scene. This is why your character looks completely different in the kitchen versus the playground.
Dataset defaults and bias. If a model's training data over-represents certain looks (often lighter skin), it will snap back toward that default unless you're explicit. Google's work on improving skin tone representation exists precisely because real products historically underperformed on darker skin tones when not built and tested intentionally.
Language ambiguity. "Brown skin" covers a massive range. So does "dark." So does "tan." And some words pull in unintended stereotypes, changing facial features, hair texture, or cultural markers even when you only wanted to specify color. Research explicitly cautions against equating skin tone with race because the connection creates problems in both directions.
The fix isn't "more adjectives." The fix is a stable naming system combined with a prompt structure that isolates base tone from lighting.

The Three Levels of Skin Tone Control for AI

notion image
Use the strongest level you actually need. More control isn't always better if it slows you down.

Level 1: Simple Names (Fast, Works for Quick Sketches)

Examples: "light beige skin," "medium tan skin," "deep brown skin," "very dark skin."
This works for quick ideation and brainstorming. But it's prone to drift when you generate multiple images or change scenes.

Level 2: Names Plus Undertone (Your Default for Most Work)

Examples: "medium tan skin with warm golden undertone," "deep brown skin with neutral undertone," "light skin with cool rosy undertone."
This is the sweet spot for most creators. You're adding a second dimension (hue), not just describing light versus dark. Research on skin color measurement argues you need more than a single "tone" axis because hue matters just as much as lightness.

Level 3: Anchor to a Scale (For Long Projects and Team Consistency)

Use a stable reference scale so everyone on your project uses the same anchor every time. This becomes critical when you're creating a 32-page children's book or a recurring character series.
The most useful modern scale for this work is the Monk Skin Tone (MST) Scale, a 10-shade scale developed by Dr. Ellis Monk and released openly for tech and research use.

10 Skin Tone Names You Can Use in AI Character Prompts

Below is a practical, creator-friendly naming set that maps to the MST's 10 anchors. These aren't "official MST names." They're promptable labels that humans can remember and that stay consistent across generations.
notion image
MST
Hex Reference
Anchor Name (Use in Prompts)
Plain Fallback
1
#f6ede4
Porcelain
very light skin
2
#f3e7db
Ivory
light skin
3
#f7ead0
Light Sand
light warm skin
4
#eadaba
Beige
light-medium skin
5
#d7bd96
Honey
medium skin
6
#a07e56
Caramel
tan skin
7
#825c43
Bronze
medium-deep brown skin
8
#604134
Mahogany
deep brown skin
9
#3a312a
Cocoa
very deep brown skin
10
#292420
Espresso
deepest brown skin

The Prompt Format That Actually Holds Up

Use this exact pattern:
skin tone: [anchor name] (MST [#]), [undertone] undertone
Examples:
  • skin tone: honey (MST 5), warm golden undertone
  • skin tone: mahogany (MST 8), neutral undertone
  • skin tone: ivory (MST 2), cool rosy undertone
Why it works: you're giving the model (1) a familiar descriptive word, (2) a stable anchor label you can repeat, and (3) a hue direction. The combination is far more stable than any single term alone.

Undertones: The Missing Half of Skin Tone Prompting

Most "skin tone name lists" fail because they only give you light-to-dark words. But two people can be the exact same lightness and look completely different because of undertone.
Use one of these five options (keep it simple):
  • Cool rosy (pink/red lean)
  • Cool neutral (cool, but subtle)
  • Neutral (balanced)
  • Warm golden (yellow/golden lean)
  • Olive (green/grey lean, often what people mean by "olive skin")
If you're unsure, pick neutral first. Then adjust based on what you see.
notion image

How to Use Skin Tone Prompts in Neolemon

This is where the naming system becomes genuinely powerful. Inside Neolemon, you can lock down skin tone once and then remix everything else.
notion image
notion image

Step 1: Create Your Base Character in Neutral Lighting

Open Character Turbo and generate your character with:
  • A plain background (light grey or simple studio)
  • Soft, diffused daylight
  • No dramatic color grading
  • No sunset, neon, or nightclub lighting
Example Character Turbo Description:
8-year-old girl, big brown eyes, round cheeks, shoulder-length curly black hair,
skin tone: mahogany (MST 8), warm golden undertone,
wearing a yellow hoodie, blue jeans, white sneakers
Example Background/Lighting Line:
simple studio background, soft diffused daylight, color accurate, neutral white balance
Why this matters: you're setting a clean "truth" for skin tone before scenes start messing with it. Research shows that skin tone perception is highly sensitive to lighting, which is exactly why you need a neutral baseline.

Step 2: Save Your "Skin Tone DNA" and Reuse It Exactly

Copy this into your notes and paste it verbatim into every prompt for this character:
skin tone: mahogany (MST 8), warm golden undertone
Don't rewrite it as "deep brown" later. Models treat synonyms as new instructions. One small wording change can introduce drift.

Step 3: Use Editors Instead of Re-Rolling the Whole Character

When you regenerate from scratch, you invite drift. When you edit from a strong reference, you anchor identity.
  • Use Action Editor for new poses (walking, sitting, jumping, waving)
  • Keep the description skin tone line unchanged every single time
This is where Neolemon shines. The tool produces draft images in seconds, not minutes. That's one of the main reasons people switch from traditional AI tools. Traditional tools are often slow, time out, and cause frustration. When users come back later, consistency is completely gone and they have to start from scratch. Our platform delivers instant speed and perfect consistency.

Step 4: When You Change Lighting, State That Base Tone Stays the Same

This single sentence is the highest-leverage addition for story scenes.
Add one line to your scene block:
keep the same skin tone as the reference character (MST 8), only change the lighting
This tells the model exactly what to preserve and what to vary.

Copy-Paste Skin Tone Library for AI Character Creation

Pick one anchor and one undertone. Paste into your character description.
notion image

Very Light (MST 1-2)

  • skin tone: porcelain (MST 1), cool rosy undertone
  • skin tone: porcelain (MST 1), neutral undertone
  • skin tone: ivory (MST 2), cool neutral undertone
  • skin tone: ivory (MST 2), warm golden undertone

Light (MST 3-4)

  • skin tone: light sand (MST 3), warm golden undertone
  • skin tone: light sand (MST 3), neutral undertone
  • skin tone: beige (MST 4), neutral undertone
  • skin tone: beige (MST 4), olive undertone

Medium (MST 5)

  • skin tone: honey (MST 5), warm golden undertone
  • skin tone: honey (MST 5), neutral undertone
  • skin tone: honey (MST 5), olive undertone

Tan (MST 6)

  • skin tone: caramel (MST 6), warm golden undertone
  • skin tone: caramel (MST 6), neutral undertone
  • skin tone: caramel (MST 6), olive undertone

Medium-Deep (MST 7)

  • skin tone: bronze (MST 7), warm golden undertone
  • skin tone: bronze (MST 7), neutral undertone
  • skin tone: bronze (MST 7), cool neutral undertone

Deep (MST 8)

  • skin tone: mahogany (MST 8), warm golden undertone
  • skin tone: mahogany (MST 8), neutral undertone
  • skin tone: mahogany (MST 8), cool neutral undertone

Very Deep (MST 9-10)

  • skin tone: cocoa (MST 9), neutral undertone
  • skin tone: cocoa (MST 9), warm golden undertone
  • skin tone: espresso (MST 10), neutral undertone
  • skin tone: espresso (MST 10), cool neutral undertone

Creative Skin Tone Descriptors from Nature

Sometimes you want prompts that feel more natural and less technical. These names work well for specific creative contexts:
Very pale and delicate: porcelain, pearl, alabaster, eggshell. These convey pale skin with soft, delicate qualities. "A princess with alabaster skin" evokes a fairytale aesthetic.
Warm lights: peach, cream, honey, wheat, champagne. Light skin with warmth. "Sunlit cream skin" creates a fair character bathed in warm light.
Medium and golden: olive, caramel, golden, toffee, amber, bronze. Mid-tones often associated with Mediterranean, Middle Eastern, or Latin American looks.
notion image
Deep browns: mocha, cocoa, coffee, mahogany, chestnut, cinnamon. Rich brown shades for deeper complexions.
Very deep: ebony, onyx, obsidian, midnight. Descriptors for the darkest skin tones. These terms often imply depth and richness, not just darkness.
Fantasy and unique: ashen (grayish, for undead characters), mint-green (for elves), sky-blue (for magical beings). If you're designing fantasy characters, you can absolutely prompt unusual skin colors. Just add enough context so the model knows you mean it.

Troubleshooting: When Skin Tone Changes Between Scenes

Use this like a debugging checklist.
notion image

Problem: Character Keeps Getting Lighter

Most common causes:
  • The new scene has bright backgrounds (snow, white walls) and the model "exposes for background"
  • You changed your descriptor language ("deep brown" became "dark skin" became "brown skin")
Fix:
  1. Reuse the exact skin tone DNA line from your base character
  1. Add: balanced exposure for face, preserve base skin tone
  1. Reduce scene adjectives that imply "bright airy high-key" unless you specify exposure

Problem: Character Looks Grey or Washed Out

Common cause: Cool lighting combined with a desaturated color grade.
Fix:
  1. Add undertone explicitly ("warm golden undertone" can counteract cool lighting)
  1. Add: healthy natural saturation, avoid desaturation
  1. Specify: neutral daylight for the base image, then reintroduce stylized lighting gradually

Problem: Skin Tone Is Right, But Ethnicity Cues Changed

This happens when you used broad identity words or the model filled in stereotypes.
Fix:
  1. Decouple skin tone from race terms (skin tone does not equal race)
  1. Describe concrete facial features instead (nose shape, lip shape, face structure)
  1. Describe hair texture directly (coily, curly, wavy, straight) instead of using ethnicity as a shortcut

Problem: "Dark Skin" Gave Only One Narrow Look

"Dark skin" is a single coarse bucket, and models often collapse diversity within it.
Fix:
  1. Switch to MST anchors plus undertone
  1. Add one more feature dimension (hair texture, facial structure)
  1. Do a 4-variation test with slightly different words, then lock the winner

Keeping Skin Tone Consistent Across Different Times of Day

Your storybook character will appear in daylight, indoors, nighttime, and sunset scenes. The trick is to keep the base tone constant while changing lighting separately.
Use this two-line pattern:
Line 1 (Identity): skin tone: ___ (MST __), ___ undertone, base tone consistent
Line 2 (Scene): lighting: ___, white balance ___, exposure balanced for face

Example: Daylight Scene

skin tone: bronze (MST 7), neutral undertone, base tone consistent
lighting: soft daylight, neutral white balance, even exposure

Example: Sunset Scene

skin tone: bronze (MST 7), neutral undertone, base tone consistent
lighting: warm sunset light, controlled exposure, preserve true skin tone

Example: Indoor Scene with Warm Lamps

skin tone: bronze (MST 7), neutral undertone, base tone consistent
lighting: warm interior lighting, white balance slightly warm, face exposure balanced
Why this works: you're explicitly separating what the skin is from what the light does. This prevents accidental recolors that make characters look inconsistent.
notion image

Full Workflow: Creating a Children's Book Character with Consistent Skin Tone

Here's the entire process using Neolemon for children's books.
notion image

1. Define Your Character DNA

Start by writing down everything that stays constant:
Name: Maya
Age: 7 years old
Skin tone: caramel (MST 6), warm golden undertone
Hair: shoulder-length black hair with curly texture
Eyes: large brown eyes with long eyelashes
Face: round face with soft cheeks
Core outfit: purple t-shirt, denim overalls, white sneakers
Distinctive feature: small birthmark on right cheek

2. Generate Your Base Image

Open Character Turbo and paste:
7-year-old girl named Maya, large brown eyes with long eyelashes, round face with soft cheeks,
skin tone: caramel (MST 6), warm golden undertone,
shoulder-length curly black hair, small birthmark on right cheek,
wearing a purple t-shirt, denim overalls, white sneakers
Action: standing, full body pose, natural smile
Background: simple studio background, soft diffused daylight, color accurate
Style: 3D Pixar-style cartoon
Generate until you get a version you love. This becomes your reference.

3. Create Your Scene Variations

Now use Action Editor to put Maya into different scenes. The key is keeping that skin tone line identical:
Scene 1: Maya at the library
7-year-old girl named Maya, large brown eyes, round face with soft cheeks,
skin tone: caramel (MST 6), warm golden undertone, base tone consistent,
shoulder-length curly black hair, small birthmark on right cheek,
wearing a purple t-shirt, denim overalls, white sneakers,
sitting at a library table, reading a picture book, fascinated expression
Scene 2: Maya in the garden
7-year-old girl named Maya, large brown eyes, round face with soft cheeks,
skin tone: caramel (MST 6), warm golden undertone, base tone consistent,
shoulder-length curly black hair, small birthmark on right cheek,
wearing a purple t-shirt, denim overalls, white sneakers,
kneeling in a garden, planting seeds, excited and curious

4. Test Across Lighting Conditions

Before committing to 32 pages, run a quick test. Generate Maya in four different lighting situations:
  1. Bright daylight
  1. Overcast/cloudy
  1. Warm indoor lighting
  1. Sunset
If her skin tone holds across all four, you're ready for production.

What Not to Do (Common Mistakes That Break Skin Tone)

notion image
Mistake
Why It Fails
What to Do Instead
Using different synonyms in each prompt
Models treat "tan," "brown," and "caramel" as different instructions
Pick one phrase, save it, paste it exactly
Forgetting undertone
You lose half the information about how the skin actually looks
Always include: warm golden, neutral, or cool rosy
Letting lighting override everything
"Golden hour" without constraints will orange-ify your character
Add: "base skin tone consistent, controlled exposure"
Using only "dark skin" or "light skin"
These are too vague and collapse diversity
Use the MST scale with specific anchors
Changing the style dramatically between scenes
Different art styles render skin differently
Lock your style choice and reuse it

Using Neolemon for Skin Tone Work

Our platform is built specifically for maintaining consistency across multiple images. Here's why that matters for skin tone:
Speed: Neolemon produces draft images in seconds, not minutes. This lets you iterate quickly when dialing in exactly the right skin tone.
Character Memory: When you use our Action Editor to create new poses, the character's identity stays locked. Your skin tone DNA travels with the character automatically.
Expression Control: The Expression Editor lets you change emotions without touching skin tone at all. Smile, frown, surprise, worry. The complexion stays exactly where you set it.
Organized Projects: Save your characters to projects and keep everything in one place. Never lose track of which prompt worked.
Multi-Character Scenes: When you need two or three characters interacting, use Multi Character to tag each one separately. Assign unique skin tones: @Character1: skin tone: ivory (MST 2), cool rosy undertone and @Character2: skin tone: bronze (MST 7), warm golden undertone.
Start with our free trial to see how the workflow feels. Twenty free credits gets you through several character designs.

Quick Reference Card

Print this or save it where you can grab it fast.
notion image

The Formula

skin tone: [anchor name] (MST [1-10]), [undertone] undertone

The 10 Anchors

  1. Porcelain (very light)
  1. Ivory (light)
  1. Light Sand (light warm)
  1. Beige (light-medium)
  1. Honey (medium)
  1. Caramel (tan)
  1. Bronze (medium-deep)
  1. Mahogany (deep)
  1. Cocoa (very deep)
  1. Espresso (deepest)

The 5 Undertones

  • Cool rosy
  • Cool neutral
  • Neutral
  • Warm golden
  • Olive

The Consistency Phrase

Add to any scene prompt:
base tone consistent, preserve true skin tone

Frequently Asked Questions

notion image

Should I use Fitzpatrick types in prompts?

You can, but it's not ideal for character design. Fitzpatrick is a dermatology phototype scale designed for UV response, not for visual representation. The MST scale was explicitly designed for broader representation in tech contexts and works better for creative work.

Does MST have official names?

MST is published as a 10-point scale, and most practical work uses numbers or swatches. This guide gives you promptable anchor names mapped to those points for human usability.

Can I just paste hex codes into prompts?

Sometimes models ignore hex codes. Sometimes they overfit weirdly. Hex is great for documenting a character bible, but language plus anchor plus undertone is usually more reliable for prompting.

How do I pick the right MST number from a reference photo?

Use a color picker (any design tool works) on a cheek area in neutral light. Compare roughly to the MST hex anchors. Don't overthink this. You're picking a stable prompt anchor, not doing medical measurement.

What about fantasy skin tones like blue or green?

Absolutely supported. Just be explicit: "mint-green skin," "sky-blue complexion," "pale lavender skin with silver undertone." The more unusual the color, the more context you need so the model knows you mean it.

Your Next Steps

notion image
  1. Pick an MST anchor from the table (1 through 10)
  1. Pick an undertone (warm golden, neutral, cool rosy, or olive)
  1. Generate your base character in neutral daylight using Neolemon
  1. Save the exact skin tone line and reuse it in every scene
That's it. You'll outperform 90% of the prompting advice on the internet just by following this system.
If you're working on a children's book, our AI Book Illustration Generator is specifically optimized for consistent characters across multiple scenes. Start with the free trial and see how fast you can get your character locked in.
Need more help? These tutorials walk through the full workflow:
Now go create something beautiful.

23,000+ writers & creators trust Neolemon

Ready to Bring Your Cartoon Stories to Life?

Start for Free