1000 Prompts in 1 Click - Dynamic Prompt Wildcards for Automatic 1111

[@OlivioSarikas]

#### Links from the Video #### Controlnet 1.1 Video: kzbin.info/www/bejne/sKOqfXidfKt9j5Y Dynamic Prompt Tool Guide: github.com/adieyal/sd-dynamic-prompts/blob/main/docs/tutorial.md My Wildcards Collection: drive.google.com/file/d/1p_V3LGU6xiOZ5IK1KflKpRRFRdT1lr_m/view?usp=sharing #### Join and Support me #### Buy me a Coffee: www.buymeacoffee.com/oliviotutorials Join my Facebook Group: facebook.com/groups/theairevolution Joint my Discord Group: discord.gg/XKAk7GUzAW

[@relaxation_ambience]

Olivio, please stop using sudden enormous zooming in your videos as it's very annoying and makes chaos. Instead use normal screen as you see. It's enough to see your mouse position on the screen and it's easier to find similar options in our screens just seeing your mouse position in the screen. Resizing pictures is ok, but when trying to explain options and jumping from normal to enormous size screen, it confuses a lot. Thank you.

[@OlivioSarikas]

@@relaxation_ambience Actually a lot of people like especially these zooms, because they can watch my videos on their tablet or smarphone when testing my tricks on their PC.

[@relaxation_ambience]

​@@OlivioSarikas Hmmm, ok then, if more people accept that, then I have to adapt to that also :) You provide very good information and I watch on PC. When you say something about options, then I go adjust the same to my SD app, then come back to your window and see enormous size letters and immediately can't understand where I'm. Then I have to pause and get back few seconds. And every time like that :) For example your friend Sebastian Kamph also zoom, but not so much, it's possible to understand in which part of the page action is going ;)

[@OlivioSarikas]

@@relaxation_ambience I get what you mean. I will try to zoom in less extreme

[@JasonVailUXGuy]

I'm usually watching these on mobile and find the zoom helpful.

[@tusumi]

Im brand new to this, I only started a couple of weeks ago. There is so much to learn, and it's overwhelming at times. I'd like to thank you Olivio. Your videos have been of great help to me! Cheers!

[@ANGEL-fg4hv]

Dude i feel so overwhelmed as well. And then he keeps adding more videos so it's like school every day. Also I'm inspired to buy a new graphics card because with my 1660ti , i can't really have fun with stable diffusion.

[@krozareq]

This is such a nice extension. Thank you for covering it. BTW, '{}' are curly braces. '' are angle brackets, '|' is a pipe. These are used a lot in programming.

[@PlasteredDragon]

I've been using dynamic prompts as well, you can also *nest* the dynamic prompt lists if you like. For example: a beautiful {brunette | redhead | blonde} woman wearing {a toga { in a roman temple | in a greek garden | on stage} | a spacesuit { in a sci-fi setting | on the moon | at a NASA training center} | a summer dress { and cowboy hat on horseback | while walking in an enchanted forest | and sunglasses on a tropical beach | at a fancy restaurant}} My first dynamic list is just a choice of hair color, but the second dynamic list chooses between 3 different costumes: a toga, a spacesuit, or a summer dress. But then for each of those costumes a second dynamic list offers three possible choices of setting, allowing for some control over where the costumes appear. You can also have empty choices in dynamic lists like so: beautiful Melvina Spinerkwitz wearing a floral-print one-piece bathing suit walking on a tropical beach { with her dog | with her twin sister | with her boyfriend | with a tiger | | | | } Because there are 8 choices in this list, but 4 of them are blank, half of the generated images just have Melvina walking alone. You can also combine these prompts with LoRAs of specific characters if you want to have the same prompt but have different characters appear in the result image like: beautiful { BCon | EWat | NPor } sunning herself on a sandy beach

[@JasonVailUXGuy]

Great tips! Thx for sharing.

[@DIYrogan]

Great tips

[@hebercloward1695]

This might be the coolest thing I've read today! Thanks!

[@blacksage81]

Thank you, out here dropping gems they are.

[@DanteJDM]

Every time I come back to Olivio to re-explain things to me again. Haha. Thanks for the clear and informative videos!

[@MisterRayYT]

I've been using this for a while, and as you say it's great for putting randomly created people or random real people into random locations - having wildcard lists of your textual embeddings and LORAs is really useful for this. Random ages, ethnicities, hats - lots to do. Or add random features to your landscapes. The one thing that I really wish it could do is interact with ControlNet to pass in random poses or depth maps from wildcard files too.

[@DIYrogan]

Bro can we connect on discord if you have extra time I need to learn this. 😊

[@SH-zg8zs]

This for sure is one of the most important and informative videos I've seen so far for Stable Diffusion. Thank you so much for the wildcard-collection and for all of your content!

[@chyldstudios]

So much cool stuff to play with! I need more time in the day!

[@IAteYourCookiez]

Same

[@OlivioSarikas]

Me too :) Thank you

[@reymortadeguzman3692]

I don't really have stable diffusion installed offline but I sure was able to use your prompts guide. Of course, they won't work exactly as projected but it has taught me ideas... lots of it... Thank You Very Much!!!!

[@akkitty22]

Thank you so much for showing this tool. I've been overwhelmed by all the great tools, this works great in combination with self made tools. This is a massive expansion of creative possibilities for virtually any artist and all kinds of work - architecture, design, games, whatever you can think of visually speaking.

[@josemuanespinto7675]

Hi @OlivioSarikas, first thank you for sharing all your discovers with us and to always teach us. I have a few words that I would like to write (even with m y bad English, I'm Brazilian) I start to learn how to use AI to create images with you and your videos, I think that the first that I saw was in last December and since then I have been seeing all videos that you record, thank you very much, and I would like to say that in the midst of so many highly competent artists like you and who record videos, I believe that the only one who is not Midjourney phobic is you, because you are always showing us memorable things that we can do with Stable Diffusion, thanks for that too. You are of a unique magnitude, because you give us everything without leaving anything behind, this demonstrates your detachment and your competence, great artists do not need to be afraid of competition for the simple fact that they are artists and unique, there is no way to replace them, nor with AI, nor with others. (Sorry for the length of the text, advertisers tend to talk too much.)

[@OlivioSarikas]

Thank you very much. That means a lot

[@Flackon]

Angle brackets (

[@seanreynolds1266]

Aww, you LEGEND! The wildcards collection is greatly appreciated Olivio! Another top video!

[@jaredbeiswenger3766]

Very cool. Was planning to post in your discord about potential resources much more simple than this, so very excited for this one.

[@Darkbolt83]

Greetings Olivio! Nice tutorial ! I'm a french newbie on SD, i love your content, I've learned so much and tested many things with your advices, results are amazing! Thank you a lot! I wonder if we will be able to put some text on the image without going on Photoshop to do a comic style board or something like that 😁

[@Milennin]

This is the stuff I've wanted from Stable Diffusion for a long time. Awesome.

[@SpiralingOut23]

Great info Olivio - you're the one-stop shop to A1111.

[@ArjenJongeling]

0:44 It's called 'Dynamic Prompts' now. I couldn't find sd-dynamic-prompts and tried to install it via 'install from URL' but the result was that I couldn't see the extension in the Web UI. I probably went to the wrong Github link. Anyhow, searching for the new name installs it correctly as shown in this video.

[@CarlosAmegos]

btw its called curly bracket {}... angle/chevron bracket is ... left one is called opening bracket {, and right is called closing bracket }... and then the separator | inside is usually called pipe or vertical bar but straight line down makes sense too

[@rinokumura9941]

Great video. I have a 9k words long prompt listing characters Characteristics and backgrounds. It is a very interesting extension. It chooses every characteristics, background, artstyle etc. At random and gives interesting results. Also, can you do a video on Multidiffusion Upscaler and Supermerger extensions. Especially Multidiffusion Upscaler.

[@avicap17]

comparta con la comunidad querido compañero

[@anotherrandomdude1022]

Ah, nice stuff! I was creating my own lib for random prompt generation when coding utility for automatic generation in a few steps, since I don't really was into this extension at a time, but it's really improved now, glad to see it

[@MainesOwn]

thanks for the wildcards-zip, this is useful not only for the purpose you showed but also to generally get better prompts

[@josea.amador5732]

Thank you very much for this valuable information that you give on your channel and for keeping us so updated on everything related to stable diffusion.

[@SheenaTigerspielt]

Oh... das Tool was ich ständig nutze ^^ hatte schon Angst ich habe was verpasst in all der Zeit wo ich meinen Rechner aktuell nicht mehr mit AI blocke ^^ Ist für mich vermutlich das wichtigste Plugin.

[@silverbullet126]

As always super helpful and now I've got a nights (or likely way more) of experimenting to do!

[@DXCommanderHQ]

Could we draw a person with some specified text on a garment / hoody that we can specify? That would be cool to get a slogan shown on the garment in the picture (I don't want to physically print the design, I just want the design to flow with the garment as if it was printed on, for instance on a photo realistic picture..?)

[@RobertJene]

thank you, the other video I saw on this was very hard to understand, your English pronunciation is getting a lot better. I am still going to roast you a little in my next video 😁

[@WantonDecay]

🤯 Awesome too! 🤩Thanks for the fantastic video and explanation. ❤

[@airazzledazzle]

that magic prompt is super handy, especially for me, i have no idea what to generate other than copy prompts from civitai

[@TheDontletmedie]

You're amazing! Thank you so much for making these videos! I'm greedy and I want MORE --thanks Olivio!

[@makesnosense6304]

It seems to me that with AI, it will be more about using different tools/models to create something than just using one tool/model. I've been thinking of generating 3d models for example, it might be easier to use controlnet to generate different types of images (depth, openpose, normalmaps), generate different angles, and then use some other tool to build 3d models out of that. Right now it's possible already, but still quite a few steps and cleanups are needed.

[@OlivioSarikas]

Yes, totally. Since we have no sufficient way to tell the AI want we want, we need to guide it by all kinds of tools and inputs and models towards what we need

[@makesnosense6304]

@@OlivioSarikas Yeah. Can't wait to see a video of you showing something like this! Also, thanks for the videos. They are quite neat.

[@c0nsumption]

@@OlivioSarikas the exciting part is that’s the first steps to creating an A.I. that can. Finding the data :)

[@MrSongib]

4:54 One tips from me if you want a colored but "natural" hair (idk how to say it. xd), use something like "[ash gray | Pink] hair" or "[blonde | pink] hair" to get natural pink hair, instead of "pink hair". the "shuffle" will make the hair color more natural, and you can put the weight inside the bracket as well, something like this "[(ash gray:1.3) | pink]". it better to use this method with sampling step above 25-50. and if you want use the mix method like the video you can do it too with something like this [ash gray | {Pink | green | blue | purple | orange}]. hope it useful.

[@patagonia4kvideodrone91]

In comfyui there is a similar one that can be put from files (which I don't use) and then another in which they are put with square brackets for example {tall|average|short}, of course the file system, if one requires much more variety it is better, due to the length of the prompt

[@unlost4z]

Really appreciate what you are doing. Hats off to you Olivio!

[@mistraelify]

Well i don't like wildcards but your chatgpt info early on the video helped me a lot didn't though about it thank you !

[@sinisastimac2264]

Olivio thanks for great tutorial and files ! Helped me a lot as beginner.

[@Oxes]

we def need a hands tutorial, for perfect hands all the time

[@makesnosense6304]

Big thanks for doing these videos. It's really helpful!

[@yuhanzhang2882]

wow it was able to keep the face constant. I thought stable diffusion would vary it every time. so much development in a few months.

[@KpMak]

Magnificent Job! Clear and concise! very useful!

[@grahamharris7010]

Thank you for the video and pre-made lists kinds Sir.

[@esdeceerde3948]

exactly what I need now! ✨ big thanks bro!

[@jessicatimm559]

Thank You for this tutorial! Very helpful.

[@lovestories259]

Thank you so much, I really appreciate your thoughtful guidance, it has helped me a lot. Once again, thank you!

[@siltoruz3502]

Hey Olivio! Great guide as always! I have been wondering for a long time, are you a Greek? Your name sounds Greek.

[@aronleniger]

Thank you for all the freat tuts ❤️

[@JorgeLuisAR]

this is amazing, where can i learn about the syntax to fully manipulate de prompts?

[@ImMikeLo]

I really like your tutorial, I am trying to find the Dynamic Prompts, but it doesn't show me anything.

[@avicap17]

Impresionante video muchas gracias saludos desde argentina

[@pol3055]

Love your videos, keep em coming

[@Aquana01]

as usual very insightful video!

[@masterprog48]

thank you, and how to wildcard of poses, I make woman with different hairstyles but she only one pose in 10 generated images

[@TamalPlays]

Can you make a updated tutorial about making lora? thank you

[@RodgerE2472]

You're a Hero!

[@IshioRamo]

Thanks again man, I needed that video XD

[@serioserkanalname499]

Didnt know you can choose multiple prompts from one curly bracket with $$. I was just chaining multiple copies of the brackets together hoping it didnt pick the same word from each until now lol.

[@Flackon]

Reading the documentation gives you superpowers!

[@Aviator-ce1hl]

I think when making a video instead to start on how to install something it would be better to start with a brief description on what you are going to talk about can do with a few examples so who is interested can then follow how to install and use it. Sometimes it happened that I installed something some video but then finding out I didn't really need it.

[@TrapWaveFusion]

Hi, thanks for this, just have a question, it can be used in collab....???

[@shiccup]

thanks for not actually showing the extension at all inside the prompt so we can see if we have the correct one... i downloaded a dynamic prompt but it didnt say sd so who knows if thats right i mean i appreciate the fancy zoom in animation but could you try to be more clear as to what to do during the actual installation part I can be a little slow but like can you actually show the thing on screen for at least 1 frame.

[@shiccup]

i did get the correct one. just making a note.

[@CuddleUTube]

sd-dynamic-prompts doesnt exist anymore??

[@jegajega3]

GREAT VIDEO 9/10 !!!!!!

[@salmon85]

wildcards can also pull from other wildcards I have one called __clothes__, and inside it is __top_clothes__, __bottom_clothes __full_clothes__, you can also use the {} shown @4:36 in the wildcard files I queued up a few thousand images overnight to see what generated.. Wasn't disappointed.

[@Braulio_Cyberyisus]

I was wishing something like thiiis Btw, can you make a video talking about regional prompt extensión?

[@yolamontalvan9502]

You’re the best. Thanks.

[@joannot6706]

he way you say "Automatic 1111" with the Audi pronunciation ^^

[@Vanguard_dj]

Dynamic prompts is awesome, though it does seem to break a lot for me.

[@warisdesigns]

I did not find the Stable Diffusion install link.... where is the "Stable Diffusion" install link?

[@blasteekmag]

Hi Olivio. Thank you for this but when i search for it it doesn't appear. However the folder exists already... weird. i was facing issues with my sd1111 and had to delete and reinstall yesterday so maybe they added it in the latest version?

[@RobertWildling]

Would there be a chance to see your complete negative prompt? The first line is unfortunately not completely visible. Thank you! 🙂

[@DIYrogan]

Hey awesome video. ❤

[@thenewsasia]

Sir we can change body style and keep dress and environment ?

[@orirune3079]

If anyone is finding that their dynamic prompts is showing up, everything seems to be loading fine, and then...it just doesn't work - check your spelling. A single misspelled filename seems to throw the entire thing into chaos.

[@jagss]

You are awesome ❤

[@OlivioSarikas]

Thank you so much!

[@roterroter9144]

Thanks a lot.😍

[@duniasainsasrofi]

mindblowing

[@AiNomadArt]

Is it possible to remove the image information of the PNG file in Stable Diffusion?

[@patrickschwenninger8120]

for me it says "No values found for wildcard [wildcardname]. Wildcards are saved as mentioned and i even restarted my whole system.

[@vl7823]

my man got drip.

[@RobertJene]

4:43 {} those are curly brackets And this characer is called a pipe |

[@XetXetable]

{ and } are actually called "curly brackets" or "braces". Chevrons, also called "angle brackets", are a different kind of bracket denoted ⟨ and ⟩.

[@Wow88z]

Hey Olivio! What’s your favorite model to use?

[@OlivioSarikas]

I will make a video about that in the next days ;)

[@JasonVailUXGuy]

revAmimated is super versatile and gives great results. But for me the model that I use depends upon the result that I'm trying to achieve. Kenshi gives beautiful anime styles. Lora's will add a ton of style guidance to augment whatever the inherent for the model. I found it helpful to generate the preview thumbnails for each of my models to give them a quick test.

[@RedRojo210]

Do they have something like this for midjourney?

[@TheDontletmedie]

at this point of the video (exactly 10:42 time stamp) where do those prompts appear? I don't see that show up in my prompt magic

[@TheDontletmedie]

Does anyone know where these prompts appear? I'm not seeing it. I'm hoping @Olivio answers me but if not --anyone else help me out?

[@shlomitgueta]

did you stop using vladumatic?

[@lavinchi6709]

SUPER👑🚀✨

[@OlivioSarikas]

Thanks 🤗

[@brianjohncuizon2872]

I somehow do not see the sd-dynamic-prompt among the list of options to install. The only one closest I see is Dynamic Prompts prompting. Any other way to install it?

[@MHBlackAce]

its only called dynamic prompts now

[@BlankFX]

I don't yet get what that could be useful for. Isn't prompting mostly obvious? I mean the prompt gets generated in our head, not randomly. If I imagine a person with an athletic body, I add "athletic" to the prompt. Why randomize all this? Isn't SD about generating AI Art as precisely as we imagine it in our mind? If I want random art, I can use Google Image Search.

[@OlivioSarikas]

For experimentation and automation. Sometimes you just want to try a lot of different hair styles or hair colors or clothing types.

[@MrClayhead]

Thanks!

[@OlivioSarikas]

Thank you so much :)

[@rajasthanelectronics]

amazing

[@TobinatorXXL]

One Button Prompt the best Script

[@jessecool561]

Does this work with SDXL as well?

[@maxziebell4013]

Is there a MJ prompt version of this?

[@CBikeLondon]

Thanks

[@OlivioSarikas]

Thank you, really appreciate it

[@lookbook.ai.]

How to train a LyCORIS (LoCon/LoHA) model?

[@ravnOne65]

Hi Olivio, it would be super cool if you could do a video on chaiNNer

[@olternaut]

Can you do me a favor please???? Can you please test Stable Diffusion on a Intel Arc A770 16GB using the OpenVino software? I'm seriously considering the GPU for use with SD.

[@JasonVailUXGuy]

Does the GPU have CUDA cores and PyTorch support?

[@blackbeans8096]

amazing!

[@Nakasasama]

Any chance on showing how to make a yaml script?

[@SheenaTigerspielt]

1:50 ah, die Idee eien andere AI für Dinge wie Haarstile zu nutzen, auf die bin ich nicht gekommen. Ich hab die alle bisher per Hand erstellt... und diverse mitgelieferte. Die wichtigsten sind aber die für z.B. Haarstile.

[@adyketobingnugroho32]

thanks dude