Skip to content

Latest commit

 

History

History
46 lines (34 loc) · 1.88 KB

README.md

File metadata and controls

46 lines (34 loc) · 1.88 KB

PI-Chan

Get prompts/descriptions of images uploaded on discord.

You can add PI-Chan into your server without hosting!

Just join the touhouai discord, and click on PI-Chan's profile to see the "Add to Server" button. Click it and you'll invite it!

Discord Server

Functionality

This Discord bot reacts to any image with generation metadata from the most/all popular webuis. (Create an issue if your favorite webui doesn't have it!) If you want to get a rough prompt (predicted with wdtagger), react with ❔

Install this if it breaks!

Setup

  1. Clone the repository
  2. Install the dependencies with pip install -r requirements.txt
  3. Create a Discord bot and invite it to your server
  4. Enable the Message Content Intent in the Discord developer portal
  5. Enable the Server Members Intent in the Discord developer portal
  6. Copy/rename the config.base.toml into config.toml
  7. Set TOKEN = "<your discord bot token>" in the .toml file
  8. So like TOKEN = "HFBVSAOa876vat764bq8967fgh8d8a76"
  9. Add the channel IDs you want the bot to work in into the config.toml file
  10. Run the bot with python3 PromptInspector.py

Examples

1 2 3

Webui Support

Webui Supported?
Automatic 1111, sd.next, forge ✅
ComfyUI 🟨*
Stable Swarm ✅
NovelAI ✅
Fooocus ✅**

*It returns the workflow and tries to extract the prompt, loras and checkpoints used

**Please test in TouhouAI, I think it will work though