Qt desktop application for building image-generation prompts with visual tag examples.
This project was created as part of MIEM project No. 1515, focused on modern image-generation technologies and prompt construction for tools such as Midjourney, DreamStudio, and Stable Diffusion.
- Provides a PySide6 desktop UI for browsing prompt tags by model and category.
- Builds a prompt by selecting tags and model-specific parameters.
- Supports manual prompt editing and copying the final prompt to clipboard.
- Loads prompt/tag structure from JSON resource files.
- Falls back to a bundled placeholder image when the full generated image dataset is not included.
- Python
- PySide6 / Qt
- JSON-based prompt catalogs
- pyperclip
This is an archived student/applied AI project. The repository was cleaned for portfolio visibility: large generated datasets, model checkpoints, training artifacts, and binary archives were removed from the public source tree.
The original project used a large set of generated example images. The cleaned version keeps the application code and compact resource files, while missing demo images are replaced with a fallback illustration.
promtManagerQt/
main.py # application entry point
prompt_widgets.py # main UI widgets and prompt-building logic
config.py # resource paths and UI constants
resource/ # JSON prompt catalogs, logos, fallback image
utility-source/
kadinskyAPI/ # sanitized API notebook example
requirements.txt
python3 -m venv .venv
source .venv/bin/activate
pip install -r requirements.txt
cd promtManagerQt
python main.pyThe Kandinsky/FusionBrain notebook is kept as a sanitized example. API credentials must be provided through environment variables:
export FUSIONBRAIN_API_KEY="your-api-key"
export FUSIONBRAIN_SECRET_KEY="your-secret-key"Do not commit real API keys, generated datasets, archives, or model checkpoints to this repository.