CivArchive
    Preview 90529539

    Brax’s Mystery Machine (NSFW + SFW VERSIONS)

    Hi sharks,

    Are you down bad with the autism, and looking for a way to randomize variables in your COMFYUI prompts?

    Well, I’ve got the tool for you!

    I’m looking for a 0% stake in my company, for 1,000,000 dollars.

    I’ve created a simple python script, called “Brax’s Mystery Machine” that loads variables from .txt files in the same directory as the script and attaches them to a prompt that is well built for most AI image / video generators.

    I know wildcards exist, and other add-ons for comfyui, but I always seem to have difficulty transitioning to another type of pose / scenario for the character.

    The current script serves as a bridge to help randomize attributes of a character (female) in different prompts. It can also be a good bridge for folks that aren’t using COMFYUI

    Here’s some things that this script currently randomizes:

    Hair (type and color)

    Skin Color

    Body Type

    Breast Size

    Eye color

    facial expression

    makeup

    Clothing (type and color) (The script can handle tops and bottoms)

    stockings / leggings

    shoes (type and color)

    Makeup

    Jewelry

    Nipple piercings (I find that nipple piercings while wearing any top makes images come out weird, so I made this setting only work if the character is topless or is lifting their top in the prompt)

    Glasses

    Facial expressions

    Pose / Scenario that the character is in

    Here are some other add-ons that can be toggled in app manually:

    Long sleeve - only pulled from tops/long_sleeve_tops.txt

    Short sleeve - only pulls from tops/short_sleeve_tops.txt

    Sweaters - only pulls from tops/sweaters.txt

    Bras - only pulls from tops/bras.txt

    Pants - only pulls from bottoms/pants.txt

    shorts - only pulls from bottoms/shorts.txt

    Skirts - only pulls from bottoms/skirts.txt

    panties - only pulls from bottoms/panties.txt

    No top (removes top type / color and replaces this with: No top, Breasts exposed, Nipples exposed)

    No bottoms (removes bottom type / color and replaces this with: no bottoms, crotch exposed, pussy exposed)

    No shoes (removes shoe type / color and replaces this with: no shoes, bare feet exposed)

    No Nipple Piercings - disables the item from being pulled into the script

    No Glasses - disables the item from being pulled into the script

    No Makeup - disables the item from being pulled into the script

    No Stockings - disables the item from being pulled into the script

    No jewelry - disables the item from being pulled into the script

    No expression - disables the item from being pulled into the script

    Don’t like a current item that’s in the mix? Remove it from the .txt file! Want more options? Add them to the .txt file!

    Have a lora you’d like to add in to a prompt / item? Add it to the respective .txt file, and it’ll always get created. I didn’t add any items / poses with lora’s, as I would hate for you to have to download a bunch of stuff you might not use.

    EXAMPLE: Lifting her top, blushing, looking away from viewer, <lora:cfg_scale_boost_854525:0.5> | top lift | glasses |

    Find a certain item / color that makes you go “WOW!”? Lock it by clicking on the title for that item, it will turn red and be immune to randomizing!

    Each one of the .txt files are where variables are stored, follow this formatting when adding / removing items and scenarios:

    1

    2

    3

    Apple

    Banana

    Horse

    ETC….

    I spent a lot of (autistic) energy on the way that poses / scenarios are handled from the pose.txt file in the main.py logic.

    Essentially, you can hardcode different modifiers for the character at the end of a pose / scenario so that it’s always handled that way.

    Here’s the modifiers:

    Topless

    Bottomless

    Shoeless

    Nude

    Top Lift

    No Makeup

    no jewelry

    No Nipple piercings

    No Stockings

    no expression

    no glasses

    With any of these predefined in your pose / scenario, it will act the same as manually clicking the checkbox in app.

    You can have multiple modifiers running on a pose / scenario at the same time, you just need to format them with a break symbol at the end of your prompts, IE:

    Sitting on a couch, legs spread, looking away from viewer | top lift | bottomless | shoeless

    I have also created pose categories, as I was getting lost trying to find certain prompts that were in the pose.txt file.

    There are some categories built out already, and you can create new ones by using [category name here], at the top of a said category in pose.txt

    Here’s the categories that I’ve already built out:

    [ALL] (Built in) (This will show all poses / scenarios for pose.txt, no matter the category)

    [MASTURBATION]

    [FLASHING]

    [POSE]

    [AFTER SEX]

    [BLOWJOB]

    [MISSIONARY]

    [DOGGY]

    [HANDJOB]

    [COWGIRL]

    The default start of any prompt piece is as follows:

    "score_9, score_8_up, score_7_up, masterpiece, best quality, newest, absurdes, pixelsketcher, very awa, newest, 1 woman"

    You can modify the main.py file so the start of your prompts will look how you’d like.

    Not a big NSFW person? No Problem! I’ve created a different version of this app that doesn’t have that stuff in it!

    Requirements / usage:

    Python 3.10+

    https://www.python.org/ftp/python/3.13.5/python-3.13.5-amd64.exe

    PYPERCLIP

    open command prompt, after python is installed, and run:

    pip install pyperclip

    After these two, just double click the .bat file and the program should launch!

    If you have any issues getting this to launch though, please let me know.

    All in all, I know this creation is a lot (super autistic), and I accept that, but I like when I can randomize things, and this… helps my sanity when I build images / videos.

    Got a question, suggestion, or just want to call me autisic AF? Bleep Bloop it down in the comments, and I’ll read / answer accordingly.

    I’m always down for looking at your creations! If you’ve modified a .txt file or the entire script, send it my way, and I’ll be sure to shout you out and post a link to your creation!

    Post your created media at the bottom of this page! I’d love to see whatever you can make!

    I will never monetize this package, or will ever reach out to you for any kind of monetary values, please be careful when talking to strangers gamers! If you take this package and try to sell it, I will blow your house up and subscribe your phone number to a bunch of scammer call centers.

    Nonetheless, have fun my discord kittens, and let me know if you need anything.

    -Brax

    V2.0 - changelog - 7/26/2025

    Added glasses as a modifiable attribute in: glasses/glasses.txt

    Added jewelry as a new item with its own .txt file: jewelry/jewelry.txt

    Added makeup as a new item with its own .txt file: makeup/makeup.txt

    Stockings automatically disable from the prompt if the item is from bottoms/pants.txt

    If a pose has a modifier it will turn that variable red and erase the text, to eliminate confusion.

    new overrides:

    long sleeve - only pulls from tops/long_sleeve_tops.txt

    short sleeve - only pulls from tops/short_sleeve_tops.txt

    sweaters - only pulls from tops/sweaters.txt

    bras - only pulls from tops/bras.txt

    pants - only pulls from bottoms/pants.txt

    shorts - only pulls from bottoms/shorts.txt

    skirts - only pulls from bottoms/skirts.txt

    panties - only pulls from bottoms/panties.txt

    Adding better prompting so images come out more slick.

    combined colors for clothing, hair and skin still separated.

    I recommend adding items to the .txt files like this: satin_skirt

    The final output will look like: (purple_colored_satin_skirt:1.0)

    Because the new formatting has so many items crammed into the prompt, I recommend turning CFG down a little, so the VAE and Model adhere to the prompt better.


    Added a setting button in the top left to change the background color in the GUI

    If you ever have a question, suggestion, or want to quote the bee movie, shoot it down in the comments so I can see it!

    Love this project and community! Thanks again for the support!

    -Brax

    Description

    V2.0 - changelog

    Added glasses as a modifiable attribute in: glasses/glasses.txt

    Added jewelry as a new item with its own .txt file: jewelry/jewelry.txt

    Added makeup as a new item with its own .txt file: makeup/makeup.txt

    Stockings automatically disable from the prompt if the item is from bottoms/pants.txt

    If a pose has a modifier it will turn that variable red and erase the text, to eliminate confusion.

    new overrides:

    long sleeve - only pulls from tops/long_sleeve_tops.txt

    short sleeve - only pulls from tops/short_sleeve_tops.txt

    sweaters - only pulls from tops/sweaters.txt

    pants - only pulls from bottoms/pants.txt

    shorts - only pulls from bottoms/shorts.txt

    skirts - only pulls from bottoms/skirts.txt

    Adding better prompting so images come out more slick

    combined colors for clothing, hair and skin still separated

    I recommend adding items to the .txt files like this: satin_skirt

    As the final output will look like: (purple_colored_satin_skirt:1.0)

    Because the new formatting has so many items crammed into the prompt, I recommend turning CFG down a little, so the VAE and Model adhere to the prompt better

    If you ever have a question, suggestion, or want to quote the bee movie, shoot it down in the comments so I can see it!

    Added a settings button in the top left that to change the background color of the GUI

    -Love this project and community! Thanks for the support!

    -Brax

    FAQ

    Other
    Other

    Details

    Downloads
    18
    Platform
    CivitAI
    Platform Status
    Available
    Created
    7/26/2025
    Updated
    4/27/2026
    Deleted
    -

    Files

    braxsMysteryMachine_sfwVersion20.zip

    Mirrors