Character Creation Act 1: Image Prompts (Making Your Bot Look Like Them)

MiocAI characters live or die on one thing: visual consistency. If your character’s face, hair, and body “reset” every time you generate an image, the roleplay starts feeling like you’re swapping actors mid-season. (baywatch..)

This first act is a firm look into appearance prompts,how to write them, how to keep them short without losing detail, how to avoid the classic “why is it doing that” mistakes, and how to iterate fast using MiocAI’s built-in testing.

0) Before We Start: The Non-Negotiables (So You Don’t Lose the Character)

MiocAI has a few hard rules that matter for image prompts:

• No real people (no photoreal clones of streamers, celebrities, exes, politicians, etc.)
• No minors / under-18 appearance (no “looks young,” “petite teen,” “loli,” etc.)
• Public characters have stricter rules than private ones (especially around explicit avatars/opening messages).
• Illegal content stays illegal even if you label it “fictional.”

Keep your prompts clearly adult, fictional, and within platform boundaries,especially if you plan to publish.

1) The Mental Model: What Your Prompt Actually Does in MiocAI

A lot of people write prompts like they’re trying to describe the entire image. On MiocAI, that’s usually the wrong target.

MiocAI images are built from two layers:

1. Appearance Prompt (Character Prompt) This is the “casting sheet.” It defines what the character looks like every time:

• hair, eyes, face features
• body type / silhouette
• outfit (or outfit system fields)
• signature accessories

2. Action / Scene Prompt (Context-Aware Image Prompt) This is generated from chat context (recent messages) and decides:

• pose, expression, actions
• framing/angle (sometimes)
• props relevant to the scene

Your appearance prompt should NOT try to do the action prompt’s job. If you cram “smirking, lying on bed, cinematic lighting, city skyline, holding sword” into the appearance prompt, you’re basically telling MiocAI:

“Do this. Every time. Forever.”

That’s how you end up with a character who’s always shot from the same angle or always holding the same object.

2) Studio vs Built-In Avatar Generation: Start With a Strong Base Image

One of the fastest ways to get consistency is to start from a clean reference image you actually like.

Recommended workflow (especially for new characters)

• Generate images in Studio (or use the prompt test images in the character editor).
• Pick a winner.
• Use that as the character’s avatar / wallpaper / reference.

Why? Studio generation tends to be better for design exploration because you’re not fighting roleplay context. You’re just designing.

Important: Your uploaded avatar does not steer image generation

Uploading a profile picture is mostly cosmetic. MiocAI doesn’t use your avatar upload as a “reference image” during normal image generation. What controls appearance is the appearance/outfit prompt (and any outfit system you’ve set up).

So if your character’s prompt says “short silver hair,” but your avatar has long black hair,the generator will follow the prompt, not the avatar.

Exception (tooling): Some flows let you turn an image into a prompt (image-to-prompt), which is useful. But that’s converting the image into text, not using the image directly as a reference every time. More on this when we actually implement this.

3) Finding the Prompt Editor (And What Those Toggles Actually Mean)

In the character creator/editor you’ll see:

• Enable Images toggle
• Style toggle: anime ↔ realistic
• Outfit prompt system (structured 4-field mode)
• Advanced model options (model picker)
• Prompt input + a Test Prompt button that generates multiple sample images

Style toggle: Anime vs Realistic

Pick one and commit early. Switching style later often requires prompt cleanup because:

• tags that work well in anime may behave weirdly in realistic
• realistic tends to punish long prompts harder

Structured prompt mode (Outfit prompt system)

This splits your appearance prompt into:

• Clothes
• Body
• Face
• Extras (often reserved / may be limited depending on implementation)

Pretty simple, and tt helps prevent your outfit and face details from stepping on each other.

4) The Golden Rule: Your Prompt Is a Permanent Tattoo

If you put something in the appearance prompt, you’re asking for it to appear in every image unless the action prompt overrides it.

That includes:

• props (“holding a microphone” → you’ll get a microphone addiction)
• framing (“full body” → you’ll keep getting full body)
• specific body parts (“bare feet” → surprise: feet forever)
• specific camera angles (“from behind” → enjoy backs)

So the question isn’t “Would this look cool?” It’s: “Do I want this in 90% of my images?”

Like what the fuck is that.

5) Prompt Format: Tags Beat Sentences (Most of the Time)

You can write natural language, but MiocAI image models generally behave better with tag-style prompts:

Good pattern

• comma-separated descriptors
• minimal filler words
• important identity traits early

Example formats

Natural-ish (works sometimes):

“A tall woman with pale skin and short white hair wearing a black jacket.”

Tag-style (usually stronger):

“1girl, pale skin, short white hair, black jacket, confident expression”

The tag version is clearer and tends to create fewer “interpretation detours.”

Here an example with natural language:

Versus with simpler tags:

6) Order Matters: Put “Identity” First, Put “Flavor” Later

Most models weight early tokens more strongly. So do this:

The recommended ordering

1. Count / subject: 1girl, 1boy, solo (optional)
2. Identity anchors: hair color/style, eyes, skin tone, signature features
3. Silhouette: tall/short, athletic/curvy/slim, wide hips, broad shoulders
4. Outfit core: main clothing items + color palette
5. Accessories: 1–3 max
6. Optional vibe tags: “elegant,” “punk,” “gothic,” etc.

What not to do (IMPORTANT)

Don’t start with a shopping list of tiny details. Your character’s identity should win the first half of the prompt.

7) Shorter = Cleaner (And Usually Higher Quality)

Long prompts don’t just “add more detail.” They also:

• create contradictions
• dilute priorities
• push your strongest features into the “ignored” zone

A practical strategy: build a 3-layer prompt

Layer A (Core Identity): must never change Layer B (Signature Look): usually consistent Layer C (Nice-to-have): optional, removed first when quality drops

Example (not for copying,just structure):

• A: 1girl, short black hair, amber eyes, beauty mark under eye
• B: black turtleneck, long coat, silver earrings
• C: subtle eyeliner, patterned scarf, rings

If images get messy, delete Layer C first.

8) Common Trap #1: “Rear Size” and Forced Camera Angles

Some body-part emphasis tags accidentally steer composition.

A classic example is explicitly emphasizing rear size,models frequently interpret that as “show it,” which nudges the camera behind the character more often than you want.

Better ways to communicate “voluptuous” without forcing the camera

Try silhouette descriptors instead of explicit focal commands:

• “wide hips”
• “defined waist”
• “curvy build”
• “full figure”
• “thick thighs” (use sparingly; it can become a focus magnet in some models)

If you want your character to read curvy but still get varied shots, silhouette terms are your friend.

9) Common Trap #2: Shoes (And Why Feet Suddenly Take Over Your Life)

If you specify footwear, many models treat it as a hint that you want to see it, which increases:

• full-body framing
• downward camera bias
• “visible feet” frequency

If you don’t care about shoes, don’t mention them

If shoes matter, use them intentionally:

• keep footwear simple (“black boots”)
• avoid extra foot-related tags unless you truly want feet visible a lot

Better framing control (without shoe obsession)

Instead of specifying shoes, control framing via:

• “portrait”
• “upper body”
• “cowboy shot”
• “close-up”

These control composition without locking you into “full body every time.”

Like what are these face feet

10) Don’t Put Props You Don’t Want in Every Shot

This is the #1 reason people get “samey” images.

If you write:

• “wearing headphones”
• “holding a sword”
• “smoking”
• “glasses” …you’re requesting those items constantly.

The accessory rule: 1–3 signature items max

Pick a small set of iconic elements and let the scene do the rest:

• one signature accessory (glasses OR earrings OR choker)
• one signature clothing element (coat OR hoodie OR dress)
• one signature feature (scar OR beauty mark OR hairstyle)

Everything else belongs in scene prompts or outfits.

11) Using the Outfit Prompt System Like a Pro (Structured Mode)

Structured mode is basically “prompt hygiene enforced by UI.”

Clothes field: keep it wearable and consistent

• Use main outfit + palette
• Don’t list 12 items
• Avoid scene-specific outfits unless you intend that outfit to be default

Good clothes field style:

• “black bomber jacket, white t-shirt, dark jeans”
• “long red dress, gold trim, black gloves”

Body field: silhouette and proportions, not anatomy fixation

• “athletic build, tall, long legs”
• “curvy build, wide hips, relaxed posture”
• “lean build, narrow shoulders”

Face field: identity anchors only

• hair style + color
• eye shape/color
• key features (freckles, scar, mole)
• keep it stable

Extras field: treat as “rare spices”

If extras is available, keep it minimal and avoid camera/scene locks:

• “soft lighting”
• “clean lineart”
• “muted palette” Not: “night city rooftop, rain, neon reflections” (that’s a scene)

12) Model Choice: Obedience vs Aesthetic (Pick Your Poison)

Different MiocAI image models have different personalities.

A common pattern you’ll notice:

• Some models are high obedience (follow prompts literally)
• Some are high aesthetic (prettier by default, but drift more)

So you choose based on what you’re doing:

When you want strict consistency

Pick the model that:

• locks hair/eyes reliably
• respects outfit items
• doesn’t “reinterpret” accessories

When you want the best-looking single image

Pick the model that:

• composes beautifully
• has strong lighting and rendering
• may freestyle details if your prompt is too long

A practical approach:

• Design phase: use obedience model to finalize the character sheet
• Art phase: switch to aesthetic model once your prompt is clean and proven

13) Iteration Workflow: How to Fix a Prompt Without Losing Your Mind

MiocAI gives you a built-in advantage: test images.

The “4-image test loop”

1. Write/adjust your prompt.
2. Click Test Prompt (or generate test images).
3. Look for:

• face consistency
• hair consistency
• outfit consistency
• unwanted recurring composition (same angle / same framing)
  4. Make one change at a time
  5. Re-test.

If you change 6 things at once, you won’t know what fixed (or broke) it.

Keep a tiny changelog

Even a simple note helps:

• v1: “hair keeps changing” → moved hair tags earlier
• v2: “too many full-body shots” → removed shoes tag, added “upper body”
• v3: “glasses disappear” → moved “glasses” earlier

14) Debugging Guide: Symptom → Cause → Fix

Symptom: Hair color changes every other image

Cause: Hair tags too late, prompt too long, model freestyling Fix: Move hair tags into the first 25–40% of the prompt; delete weak extras

Symptom: Outfit becomes inconsistent

Cause: Outfit list is too detailed or contradictory Fix: Reduce outfit to core items + palette; move “fashion details” to outfit system/outfits

Symptom: Always full-body

Cause: Shoes, “full body,” overly detailed clothing, body-part focus tags Fix: Remove footwear; add framing tag like “portrait” or “upper body”

Symptom: Always same angle

Cause: Angle tags in appearance prompt (“from behind,” “from below,” etc.) Fix: Remove angle tags from appearance prompt; use scene/action prompts instead

Symptom: Random props appear

Cause: Vague tags that imply props (“idol,” “performer,” “soldier”) Fix: Replace vague role tags with visual traits (“stage outfit” vs “performer”)

15) Example Prompts You Can Learn From (Don’t Copy Blindly,Steal the Structure)

Below are example prompt styles designed to show structure. Swap details to fit your character.

Example A: Minimal “stable identity” prompt (anime)

1girl, short auburn hair, green eyes, pale skin, freckles, black turtleneck, long grey coat

Why it works:

• short
• identity first
• outfit simple

Example B: Balanced prompt with one signature accessory

1boy, messy silver hair, blue eyes, small scar on cheek, slim build, white shirt, dark vest, thin tie, round glasses

Why it works:

• scar + glasses are strong identity anchors
• still not overloaded

Example C: Structured mode example (good separation)

Clothes:

black hoodie, dark cargo pants

Body:

athletic build, broad shoulders, relaxed stance

Face:

short black hair, grey eyes, sharp jawline

Extras (optional):

soft lighting

Example D: “Curvy without forced framing”

1girl, long brown hair, hazel eyes, curvy build, wide hips, fitted sweater, high-waisted skirt, hoop earrings

Why it works:

• silhouette term (“wide hips”) instead of “focus on body part”
• outfit readable, not encyclopedic

📷 Image placeholder: “Curvy silhouette with varied camera angles.”

16) Advanced Tips (Use Only After You’re Getting Consistent Results)

A) Emphasis syntax (if supported by your model/tools)

Some systems support strengthening/weakening tokens (often {} or []). If your model respects it, it can help when one detail keeps getting ignored.

Use sparingly:

• too much emphasis makes the image “clingy” to that detail

B) Quality boosters (optional)

Some models respond to generic “quality” tags. They’re not magic, but can help sometimes:

• “best quality”
• “high quality”
• “very aesthetic”

Tradeoff: they consume prompt space and can be placebo on some models. Test once,keep only if it clearly improves results.

17) The “Do / Don’t” Cheat Sheet

Do

• Put identity traits early (hair, eyes, signature features)
• Keep the appearance prompt focused on what should persist
• Use structured mode if you like clean control
• Use the test images loop and change one thing at a time
• Prefer silhouette descriptors over “camera bait” anatomy tags

Don’t

• Add props you don’t want forever
• Add shoes unless you want more full-body / feet visibility
• Add camera angles to the appearance prompt
• Write a 400-character prompt and expect stability
• Expect avatar uploads to influence generation automatically

18) What Act 2 Will Cover (So You Don’t Misuse Act 1)

Act 1 is about designing a stable character sheet.

Act 2 will focus on personalities. thats all for today, take care!