Automatically correct typos, capitalization, grammar errors, and casing using a large language model, accessible via convenient keyboard shortcuts.
- Offline AI Correction: Leverages a local large language model (LLM) for real-time text correction, no internet connection is required.
- Hotkey Integration: Trigger text correction seamlessly using F9 (for selected text) and F10 (for the current line).
- Clipboard Interaction: Copies corrected text to the clipboard for effortless pasting.
- Customizable: Potentially adapt to different LLM endpoints and models for specific needs.
Ensure you have Python and the required libraries installed:
pip install -r requirements.txt
Set up a supported local LLM endpoint (see notes for configuration).
- Download Ollama: https://ollama.com/
- Run:
-
ollama run gemma:2b-instruct-q4_K_S
- Run the script:
python main.py
- Use the keyboard shortcuts:
- F9: Correct the currently selected text.
- F10: Correct the entire line of text where your cursor is positioned.
Key Libraries:
-
pynput: Listens for keyboard events and triggers text correction.
-
pyperclip: Interacts with the clipboard to copy and paste text.
-
httpx: Communicates with the local LLM endpoint for text processing.
-
logging: Logs events and potential errors for debugging.
-
LLM Configuration:
- Ollama REST API
- Default model: gemma:2b-instruct-q4_K_S
.
Abdulmunim Jundurahman Jemal (https://github.com/abdulmunim-jemal)
- Replace any placeholder LLM information with your specific setup. (Explore Ollama.com for more details.)
- Consider adding more detailed configuration instructions for users unfamiliar with LLMs.
- Explore incorporating a visual interface or additional features to enhance user experience.