Introduction
SD-WEBUI-DISCORD is a Discord bot developed in Go language for stable-diffusion-webui. It utilizes the sd-webui-go to invoke the sd-webui API and supports cluster deployment of multiple sd-webui nodes with automatic scheduling and allocation. At the same time, there is also the sd-webui-discord-ex, which is an extension on the stable-diffusion-webui that you can install and use directly. It will automatically update every time you restart SD webui.
Screenshots
News
2023-09-24: Support multi image generate in txt2img
command
2023-09-23:
- Support
Retry
andDelete
intxt2img
command
2023-09-22:
- Support
setting_ui
command - Better
txt2img
response ui,thanks for venetanji support that! #5 - Optimize command add when bot start,thanks for venetanji support that! #5
2023-09-10: Support local language
2023-09-05: Support User Center on Windows
2023-09-04: Support Image to Image
2023-08-31: Support User Center
2023-08-27:
-
Support txt2img choice model checkpoint
-
Support upload image with attachment:
deoldify
png_info
roop_image
commands
2023-08-26: Support Setting default sd-webui options
2023-08-23: Support ControlNet for txt2img
command
- By using the
controlnet_detect
command to obtain the parameters of ControlNet and filling them into thecontrolnet_args
parameter of thetxt2img
command, you can use ControlNet in txt2img.
2023-08-22: Support txt2img
command
2023-08-22: Support roop
command
2023-08-20: Support controlnet_detect
command
Features
- Local language support.
- Global sd-webui default options.
- Supports multi-node (sd-webui) deployment, distributed cluster queue with automatic scheduling.
- User Center
- Set user options with ui.
- User Center can be freely enabled.
- Customizable database types, currently supporting MySQL and SQLite.
- Users can set their own default options. If not specified when generating images, the user's configured options will be used by default, such as image dimensions, cfg, steps, etc. For more details, please refer to the User Center (wiki).
- Supports user registration.
- ControlNet Preview
- Supports specifying module and model through selection, eliminating the need for manual input.
- Allows previewing preprocessing effects and obtaining args simultaneously (used for user-generated txt2img).
- Text to Image
- Support SDXL's refiner!
- Supports user-defined default options from the User Center.
- Enables specifying model, sampler, and other optional parameters through selection, eliminating the need for manual input.
- Supports using parameters from "ControlNet Preview" directly.
- Image to Image
- Allows uploading images through an image control, rather than using URLs.
- Supports all operations of img2img in sd-webui!
- Roop Face Swap
- Allows uploading images through an image control, rather than using URLs.
- Supports specifying source and target through selection, eliminating the need for manual input.
- Supports custom face rendering algorithms (GFPGAN, CodeFormer), as well as weights.
- Deoldify Colorization
- Allows uploading images through an image control, rather than using URLs.
- Currently the best photo colorization model.
- Segment-Anything
- Allows uploading images through an image control, rather than using URLs.
- Supports image segmentation based on natural language descriptions (DION+SAM).
- Enables specifying DION and SAM models through selection, eliminating the need for manual input.
- Background Removal
- Allows uploading images through an image control, rather than using URLs.
- Supports commonly used background removal algorithms.
- Supports returning the mask.
- Extra Single
- Allows uploading images through an image control, rather than using URLs.
- Supports facial repair.
- Supports super-resolution, with models available for direct selection without manual input.
- Png Info
- Allows uploading images through an image control, rather than using URLs.
- Supports retrieving parameters of images generated by sd-webui.
Usage
The command is still under active development, and there are two ways to experience sd-webui-discord:
- Join our Discord Server where you can try out the latest features and contribute by submitting issues and pull requests.
- Self-deploy it to have your own sd-webui-discord instance.
Discord Server
Installation
You need to install the following extensions on the SD webui:
1.Download the latest release from here.
2.Create a bot account on Discord and get the token. How To Create Discord App.
3.Configuration and Startup
tar -zxvf sd-webui-discord-release-v*.tar.gz # unzip the release package
cd sd-webui-discord-release-v*/release/
Edit the config.json
file and fill in the token and other information.
{
"sd_webui":{
"servers":[
{
"name":"webui-1",
"host":"127.0.0.1:7860",
"max_concurrent":5,
"max_queue":100,
"max_vram":"20G"
}
]
},
"discord":{
"token":"<your token here>",
"server_id":"<your servers id here if empty all servers>"
}
}
If you want set default value with sd-webui
{
"sd_webui":{
"servers":[...],
"default_setting": {
"cfg_scale": 8,
"negative_prompt": "bad,text,watermask",
"height":1024,
"width":1024,
"steps":32,
"sampler":"Euler",
"sd_model_checkpoint":"sd_xl_base_1.0.safetensors [31e35c80fc]"
}
}
...
}
If you want to enable the User Center
{
...
"user_center":{
"enable":false,
"db_config":{
"type":"sqlite", // support mysql and sqlite
"dsn":"./user_center.db"
}
}
...
}
If you want disable return info in img2img
and txt2img
{
...
"disable_return_gen_info":true
...
}
Start The Bot
# if you can't connect discord,you need use proxy and run this command:
# export https_proxy=http://127.0.0.1:8888;export http_proxy=http://127.0.0.1:8888;
./sd-webui-discord
Participating
This is an ongoing project, and if you are interested in contributing, you can join our Discord Server. We welcome any feedback or suggestions, so feel free to submit an issue.