A discord bot built to interface with the InvokeAI fork of stable-diffusion.
This branch is a work in progress for a major rewrite of the arty project.
It still has a long way to go before it's ready for public use and should be considered an alpha test at best.
It's currently compatible with invokeai 4.2.0
.
Working:
- Oldschool
!dream prompt
+ parameters - Supports multiple Invoke3 backends on local network (no direct file access required)
- Building node graphs from job requests,submitting,tracking,posting to discord
- Refresh button starting renders using png metadata alone (no job db required!)
- Input images can be used as sources for image to latent, controlnet, ip_adapter
- Tweak menu with models, aspect ratio, scale, steps, sampler, strength
- Remove background using custom invokeai node
- websocket job progress tracking, discord status updates for invoke cluster queue info
- LLM integration (/chat , /describe , prompt improvement)
Setup:
git clone -b arty2-invoke4-WIP https://github.com/ausbitbank/stable-diffusion-discord-bot/
cd stable-diffusion-discord-bot
mv .\config\config.json.example .\config\config.json
- Edit config.json, you need at least a
discordBotKey
,adminID
, to set a default model and check the cluster url - If you want to run the bot in docker, modify config.cluster.url to
http://host.docker.internal:9090
Install custom invokeai nodes for advanced functionality
- Enter your the nodes folder within your invokeai install
cd invokeai\nodes
git clone https://github.com/gogurtenjoyer/nightmare-promptgen
git clone https://github.com/blessedcoolant/invoke_bria_rmbg
git clone https://github.com/mickr777/textfontimage
git clone https://github.com/helix4u/interrogate_node
- Enter the "invokeai developers console" or enter the python venv for invoke and
pip install clip-interrogator
Launch natively:
npm install
npm start
OR Launch with docker:
docker-compose up --build
If you have any questions you can find me (ausbitbank) in
You can test out the bot in any of the #artspam channels or by DM'ing