A powerful Discord bot that interfaces with InvokeAI, bringing advanced AI image generation capabilities to your server.
This branch represents a major rewrite of the Arty project and is currently in alpha stage. It's compatible with invokeai 5.3.0
but is not yet ready for public use.
- Classic
!dream prompt
command with parameter support - Multi-backend support for Invoke over local networks
- Advanced node graph generation and job tracking
- Refresh functionality using PNG metadata
- Versatile input image handling (image to latent, controlnet, ip_adapter)
- Customizable settings menu (models, aspect ratio, scale, steps, sampler, strength)
- Background removal using custom InvokeAI node
- Real-time job progress tracking via WebSocket
- Discord status updates for Invoke cluster queue information
- LLM integration for chat, image description, and prompt improvement
- Node.js
- npm
- Git
- Docker (optional)
-
Clone the repository:
git clone -b dev https://github.com/ausbitbank/stable-diffusion-discord-bot/ cd stable-diffusion-discord-bot
-
Set up the configuration:
mv .\config\config.json.example .\config\config.json
Edit
config.json
and set at leastdiscordBotKey
,adminID
, and check the cluster URL. -
Install custom InvokeAI nodes:
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
-
Install additional dependencies:
pip install clip-interrogator
npm install
npm start
docker-compose up --build
Patches and pull requests are greatly appreciated! Feel free to contribute to the project.
If you have any questions, join our Discord server or DM the bot directly.