Sora RPA Automation

Desktop automation scripts for a Sora/ChatGPT image and video prompt workflow. The scripts use PyAutoGUI image matching, clipboard automation, and local prompt files to paste prompts, copy generated output, parse option blocks, and click through image/video creation screens.

Features

• Uses PNG reference images to find buttons and UI locations on screen.
• Copies master prompt text into the clipboard.
• Captures generated text from the clipboard into local output files.
• Parses OPTxx options and image prompt sections from generated text.
• Automates repeated UI steps for creating and saving image/video assets.
• Supports a configurable working directory through SORA_RPA_DIR.

Requirements

• Python 3.10+
• Python packages from requirements.txt
• A visible desktop session with the target browser/UI open
• The included PNG reference images must match your current UI layout, zoom level, and display theme.

Configuration

By default, scripts read and write prompt/output files in this folder. To use another working directory, copy .env.example to .env or set SORA_RPA_DIR in your shell.

Expected prompt/output filenames include:

• MASTERPROMPTWHICHX.txt
• MASTERPROMPTCINEMATICINTRO.txt
• OUTPUTA.txt
• OUTPUTB.txt

OUTPUTA.txt and OUTPUTB.txt are generated files and are ignored by Git.

Usage

Install dependencies:

pip install -r requirements.txt

Run the specific automation script you want from this folder, for example:

python "Basic Controller 2 Final.py"

Do not use the mouse or keyboard while a script is controlling the desktop.

Safety Notes

These scripts automate live desktop actions and depend on screen coordinates and image matching. Test each workflow slowly after any UI, browser zoom, display scaling, or image-reference change.

Built With

• pyautogui
• python
• sora