OneButtonPrompt
One Button Prompt
Install / Use
/learn @AIrjen/OneButtonPromptREADME

One Button Prompt
Summary
One Button Prompt is a tool/script for Automatic1111/ComfyUI/RuinedFooocus for beginners who have problems writing a good prompt, or advanced users who want to get inspired.
It generates an entire prompt from scratch. It is random, but controlled. You simply load up the script and press generate, and let it surprise you.
It is a full AI prompt generator for Stable Diffusion.
It is best used on all-purpose models, such as Stable Difussion 1.5 or SDXL type models. However, feel free to use it on your personal favorite models.
A simple user guide for first time use and settings is available here.
It is also now available as a custom node for ComfyUI. Check installation doc here
More of a Fooocus fan? Take a look at this excellent fork called RuinedFooocus that has One Button Prompt built in. Check some options available here
Any other AI tool you are using? Midjourney? Dalle? No problem, I got it working on a website here. Just copy the prompt to your clipboard with a click, and paste it in any image generator tool.
Features
- Full prompt generation with the click of a button. ==> guide to my first generation
- Supports TXT2IMG, IMG2IMG, ControlNET, inpainting and latent couple. ==> guide to IMG2IMG and ControlNET
- Save your favorite generation settings with presets. ==> One Button Presets
- Workflow assist, generate multiple prompts with One Button. Create prompt variants with ease. ==> guide to workflow assist and prompt variant mode
- Create infinite variations of a chosen subject. ==> guide to override subject
- Fully automated generation, classification and upscaling. ==> guide to one_button_run_and_upscale
- Add LoRA's, customize lists and personal artists choices. ==> guide to custom files
- Use Anime Model mode to generate prompts specific for use with anime/pony models ==> guide to Anime Mode
- Compound multiple prompts together for unexpected results. ==> guide to prompt compounder
- Use wildcards, or combine with Dynamic Prompts extension ==> guide to using wildcards and Dynamic Prompts
- Has a set of template prompts from various sources, fully wildcarded and usable with Subject Override ==> guide to prompt templates
- Has multiple prompt generation modes to choose from ==> guide to prompt generation modes
- Integration is available with the superprompt-v1 model ==> guide to super prompt
- Stay in control, and fine-tune One Button Prompt with a config file ==> config file
- Flufferize your prompt to enhance your results with Prompt Magic
Important Notice
One Button Prompt had a recent large change in the UI when Presets were introduced. This might be confusing for long time users.
If you are not finding the settings, change the "One Button Preset" to "Custom...".
<img src="https://github.com/AIrjen/OneButtonPrompt/assets/130234949/6d821c82-edf9-4496-867f-ac01dda994cd.png" width="50%" height="50%">Subtypes are also completely gone now, and integrated into the subject field.
Loading the first time after upgrading on ComfyUI might give some errors, because the fields have shifted. Just set the settings back.
How to use in automatic1111/SD.next
In TXT2IMG or IMG2IMG, select the script "One Button Prompt".
<img src="https://github.com/AIrjen/OneButtonPrompt/assets/130234949/b43f7bef-533d-4139-a477-562a0b2d39ca.png" width="50%" height="50%">Leave the prompts empty:
<img src="https://user-images.githubusercontent.com/130234949/230793068-d38bc782-4c2f-4268-9e91-76f4eabe3eca.png" alt="who needs prompts anyway" width="50%" height="50%">Hit Generate!
<img src="https://user-images.githubusercontent.com/130234949/230793086-cedbe72a-e1eb-46e5-a425-4a52540847f6.png" alt="click!" width="30%" height="30%">Enjoy creating awesome pictures: <img src="https://github.com/AIrjen/OneButtonPrompt/assets/130234949/aa0ff559-03a9-49be-bbe5-94789f8102a0.png" alt="wow, good job you!">
stylized by Waterhouse, John Constable, Ed Blinkey, Atey Ghailan, Studio Ghibli, Jeremy Mann, Greg Manchess, Antonio Moro and makoto shinkai, landscape of a Atmospheric (Lazarus Labs:1.1) , from inside of a Furious Stargate, Ultrarealistic, extremely hyper aesthetic
Please be aware, that not each picture will be awesome due to the randomness of the prompt, artist and model used. You might get an epic landscape, or a photo of an Aggregavated Trout. In my experience, about 1 in 5 is good. Everyone of them is interesting.
Don't get overwhelmed by the options, they will become more clear once you use it more.
For first time users, play around with the set presets.
Some more examples below. And check the first time user guide here.
Some details
It will generate between 0 and 3 artists, and add those the prompt.
It can generate the following subjects, while building a prompt:
-
object - Can be a random object, a building ,a vehicle, some food or objects from space.
-
animal - A random (fictional) animal. Has a chance to have human characteristics, such as clothing added.
-
humanoid - A random humanoid, males, females, fantasy types, fictional and non-fictional characters. Can add clothing, features and a bunch of other things.
-
landscape - A landscape or a landscape with a building.
-
concept - Can be a concept, such as "the X of Y", or an historical event such as "The Trojan War". It can also do a line from a poem or from a song.
It mixes techniques such as prompt switching and hybrids.
This generator will generate a complete full prompt for you, based on randomness. You can increase the slider, to include more things to put into the prompt. Recommended is keeping it around 3-7. Use 10 at your own risk.
There are a lot of special things build in, based on various research. Just try it, and let it surprise you.
Suggestion is to leave the prompt field empty, anything here will be added at the end of the generated prompt.
It doesn't add anything to the negative prompt field, so feel free to add your favorite negative prompts here.
For each Batch you run, it will create a new prompt. For each batch size, it will reuse the same prompt.
Installing in automatic1111
One Button Prompt can be found in the normal installation list of Automatic1111. Go to Extension -> Available and press Load From. In the list you will see One Button Prompt, and press install.
If you want to do it manually,sSimply install this from via install from URL. After that you can see OneButtonPrompt in the script fields for txt2img and img2img. Set the following URL to install from: https://github.com/AIrjen/OneButtonPrompt
<img src="https://user-images.githubusercontent.com/130234949/230917712-60a3c1f4-fea6-4ecd-bc68-da52f1ff86fe.png" alt="Easy to install" width="50%" height="50%">You can also download the files from github directly, and place them under your automatic1111 installation in the \Extensions\ folder.
Main tab
The main tab will show preset options, which are preset generation settings which will help achieve good results. On default, it is set to Standard, which was the previous default setting of One Button Prompt. It is quite random and give some wild results.
If you are looking for more guided results, choose a different preset. Some good first choices are:
- Standard -> One Button Prompts default. Wild and unpredictable. Good for prompt exploration.
- Unique People Portraits -> Will focus on generating photographic people
- D&D Style Portraits -> Will focus on generating fantasy style portraits
- Consistent Results -> Running on a large set of build in styles, to get consistent results in prompts. Works best on SDXL.
- Greg Mode - The Preset -> Focusses on creating characters, mixed with popular artists (like Greg Rutkowski). Works best on SD 1.5 models.
If you want to change the settings yourself, p
Related Skills
node-connect
349.0kDiagnose OpenClaw node connection and pairing failures for Android, iOS, and macOS companion apps
frontend-design
109.4kCreate distinctive, production-grade frontend interfaces with high design quality. Use this skill when the user asks to build web components, pages, or applications. Generates creative, polished code that avoids generic AI aesthetics.
openai-whisper-api
349.0kTranscribe audio via OpenAI Audio Transcriptions API (Whisper).
qqbot-media
349.0kQQBot 富媒体收发能力。使用 <qqmedia> 标签,系统根据文件扩展名自动识别类型(图片/语音/视频/文件)。
