Qwen Image Bot

A Discord bot for generating and editing images using the Qwen Image Studio API.

Features

• Image Generation - Generate images from text prompts using natural language or slash commands
• Image Editing - Edit images by uploading and describing changes
• Re-editing - Reply to bot-generated images to iteratively refine them
• Automatic Image Resizing - Large images are automatically resized (max 1024px on longest side) before upload to prevent server OOM errors
• Job Queue - View queue status and check on individual job progress
• Server/Channel Restrictions - Optionally restrict bot to specific servers or channels

Setup

1. Install dependencies:

   pip install -r requirements.txt

2. Create a Discord bot at https://discord.com/developers/applications

   • Create a new application
   • Go to "Bot" and create a bot
   • Copy the token
   • Enable "Message Content Intent" under Privileged Gateway Intents

3. Configure environment:

   cp .env.example .env
   # Edit .env with your Discord token and API URL

4. Invite the bot to your server with this URL (replace CLIENT_ID):

   https://discord.com/api/oauth2/authorize?client_id=CLIENT_ID&permissions=277025467456&scope=bot%20applications.commands
   

5. Make sure Qwen Image Studio is running, then start the bot:

   python bot.py

Usage

Natural Language (Chat)

Generate an image:

draw a cat sitting on a rainbow

Edit an image: Upload an image and include your edit instructions in the message:

[attached image] make the sky purple

The bot will reply "Got it, I have enqueued your request" and ping you when done.

Slash Commands

Command	Description
/generate	Generate an image with full options
/edit	Edit an uploaded image with full options
/status	Check the status of a job
/queue	Show current queue status
/system	Show system/GPU information

/generate options

• prompt - Description of the image to generate (required)
• negative_prompt - What to avoid
• width - Image width (default: 1024)
• height - Image height (default: 1024)
• steps - Inference steps (default: 20)
• cfg - CFG scale (default: 4.0)
• seed - Random seed for reproducibility

/edit options

• image - Image attachment to edit (required)
• prompt - Edit instructions (required)
• negative_prompt - What to avoid
• steps - Inference steps (default: 8)
• cfg - CFG scale (default: 4.0)
• seed - Random seed for reproducibility

Environment Variables

Variable	Description	Default
DISCORD_TOKEN	Discord bot token (required)	-
API_BASE_URL	Qwen Image Studio API URL	http://localhost:8000
ALLOWED_GUILDS	Comma-separated guild IDs to allow (empty = all)	-
ALLOWED_CHANNELS	Comma-separated channel IDs to allow (empty = all)	-
LOG_LEVEL	Logging level (DEBUG, INFO, WARNING, ERROR)	INFO
JOB_TIMEOUT	Job processing timeout in seconds	1000
DEFAULT_GENERATION_STEPS	Inference steps for natural language generation	10
DEFAULT_EDIT_STEPS	Inference steps for natural language edits	10
MAX_IMAGE_DIMENSION	Max image dimension before auto-resize	1024