10 Prompts for Facial Details and Poses
Eye Bags (Aegyo Sal)
The under-eye fullness reads clearly here. "large eye bags" can be misread as dark circles, but with a facial close-up it renders as aegyo sal.
1 woman,large eye bagsMole
Several small moles on the cheek and around the eyes. "mole" barely responds to position control, so generating more variations beats trying to pin down the exact spot.
1 woman,moleFreckles
Freckles scatter evenly across the cheeks. "freckles" competes with skin-texture words, so leaving out "smooth skin" style terms helps them survive.
1 woman,frecklesNo Makeup
A bare-skin texture with a hand resting on the face. "no makeup" specifies bare skin rather than plainness, so softening the lighting terms brings it out.
1 woman,no makeupTongue Out
An anime-style shot with the tongue clearly out. "tongue out" lands more reliably in 2D styles than in photoreal ones.
1 cute girl,tongue outSplit Tongue
You can see the tongue tip forked in two. "split tongue" is a fine-grained word that gets washed out on its own, so writing it together with "tongue out" — locking in the tongue-out state first — is the condition for success.
1 woman,split tongue,tongue out,smilingSalute
A clear salute with the hand at the cap brim. "salute" easily leaves the arm ambiguous, so adding a clothing term that includes a peaked cap — "military or naval uniform" — pins down the hand position.
1 cute girl, beautiful, masterpiece, extremely fine and beautiful,((military or naval uniform)),salute,short hair,wavy hair,blonde hair,blue eyes,(((ruins)))Spread Legs
A man in a suit sitting with legs apart on a sofa. "spread legs" vanishes in an upper-body crop, so a seated pose and a full-body angle are prerequisites.
1 good looking guy,spread legs,suitWet Body
A uniformed figure turning back at a rainy waterside. That said, the wetness is subtle — even in brackets, "wet body" is a weak term. Pairing it with "rain" is essentially mandatory.
((from behind)),short hair,(((wet body))),((rain))Hugging
A woman hugging a man outdoors. Multi-person scenes break easily, but stating who does what to whom in word order — "woman hugs man" — stabilises it.
woman hugs man,smilingThree Moves to Keep a Fine-Grained Word
As a rule, the finer the word, the more easily it gets washed out — aegyo sal and split tongue are typical. When it does not appear, switch tactics in this order.
"split tongue" was written with "tongue out" so the tongue came out first. Fine words survive when you build the base with a broader word and add the detail on top.
If the salute won't land, add "military uniform"; for a wet body, add "rain". Giving a scene where the action or state occurs raises accuracy.
Tongue-out was more stable in a 2D style. A word that fails in photoreal sometimes lands once you change the style. Try it in the prompt generator.
Volume one is here. More examples in the image list.
Frequently Asked Questions (FAQ)
Why do fine-grained words get washed out?
Because words with less training data lose out to stronger neighbouring words. Placing a broad base word first and adding the fine word on top makes it land more often.
Why are some examples male?
Because trait words do not depend on gender. Here the spread-legs and hugging examples feature men; you can reuse the same words by swapping only the subject.
![10 AI Image Prompts for Face & Body Traits [Free, With Examples]](/storage/ai_imgs/20231231191905_2763390.webp)
![15 AI Beautiful Woman Image Roundup [Free]](https://ja.ai-img.net/storage/ai_imgs/20231111162641_5677614.webp)
![26 Pretty Girl AI Image Roundup [Free]](https://ja.ai-img.net/storage/ai_imgs/20231111155552_8882368.webp)