Image prompting keyword catalogue

This post acts as a community made catalogue of keywords.
If you find something we don’t have yet leave a comment. Of course your findings are the most welcome

For now the whole catalogue is made with Anime-original in mind. Keep in mind that some keywords work very differently with certain models,

==============

Character physical tags

Keywords that define the look of your character
Character physical tags catalogue

==============

Backgrounds

Keywords that define the look of the background
Background tags catalogue

==============

Clothing

Keywords that define the clothing of a character
Clothing tags catalogue

==============

Poses & expressions

Keywords that help you pinpoint the pose of a character
Character poses & expressions tags catalogue

==============

Perspective

Keywords that change the camera angle
Use (Keyword):2 for the best result
Perspective Keywords Catalogue

==============

Lighting

Keywords that change light in the scene
Those are the results with (x):2
Lighting keyword catalogue

==============

Quality

Keywords that refine the picture
Quality keyword catalogue

==============

Style

Keywords that are used to achieve certain styles

==============

NSFW

Sex poses, accessories, fetishes

Roadmap:
Stuff left to do:

Backgrounds
Furries because we all want to mo… PET! pet our furry friends
Style is left as last because it will be done alongside the advanced guide to image styles.

Credits:
@Fuka - owner of the post (poses overall)
@Froyst (haircuts)
@kelsier (researching perspective)
@Ryan_2 (correcting and adding some poses, expressions: eyes)
@sanctumsempra (custom races)
@SpiritCrystal (requested Race/Face/Age categories in physical)

9 Likes

Furry
Kemono

Thanks for the catalogue. I have a question though, how to specify the keywords for one character only, in order to prevent the swap between the two ? What is MAIN_PERSONA exactly, and how to use it ? Is there a SECONDARY_PERSONA keyword too ?

MAIN_PERSONA tells model to use the physical prompts of the kyn. (unless Fuka says otherwise. When it comes to the images always listen to Fuka first)

1 Like

This question is better under the guides than catalogue.

__MAIN_PERSONA__ Can be used only with ADDBASE syntax command in SD models. It is used to inject the persona prompt after everything you wanted to add before it when using ADDBASE it adds physicals and clothing to the prompt.

Example:
ADDBASE (anime cartoon style, thick lineart, simplistic)1.3, BREAK (forest fight scene)1.2, (epic, cinamtic, shonen anime fight)1.2, BREAK __MAIN_PERSONA__ [here goes rest of the prompt]

To better understand how the image generation sees it:

The usual prompt pipeline: Image prompt + Physical + Clothing + Location

ADDBASE pipeline: Location + [Everything before __MAIN PERSONA__] + __MAIN_PERSONA__(Physical + clothing ) + everything after.

There is no second persona etc for SD because the pipeline and models do not support multi-character prompting.
ADDBASE was created to split user prompts so that you can separate parts of your prompts to transform it into regional prompting (Hence the Left/Middle/Right that you can add). With that separation you can somewhat prompt in another character into the picture.

ADDBASE [Scene description], BREAK [characters in the scene], BREAK [Character 1], (side), BREAK __MAIN_PERSONA__, (side)

1 Like