BZTyping is a powerful, customizable speech-to-text application that converts your spoken words into typed text in real-time. It features a sleek, modern user interface with a minimalist mode option and utilizes the state-of-the-art Whisper large language model for accurate transcription.
- Real-time Dictation: Speak naturally and have your words transcribed instantly
- Modern UI: Sleek, transparent interface that stays out of your way
- Minimalist Mode: Collapse to a tiny bar when you need more screen space
- Global Hotkeys: Control recording without switching applications
- Language Support: Transcribe in multiple languages
- Customizable Settings: Configure hotkeys and preferences to your liking
- GPU Acceleration: Utilizes CUDA for faster transcription when available
- Windows 10/11
- Python 3.9 or higher
- NVIDIA GPU with CUDA support (recommended for best performance)
- Administrator privileges (required for global hotkey functionality)
- Microphone
- Clone or download this repository
- Run the
install_dependencies.batscript to install all required packages - Ensure ffmpeg is installed and accessible in your system's PATH
If the automated installation fails, you may need to manually install some components:
- PyAudio often requires manual installation on Windows
- CUDA drivers may need separate installation for GPU acceleration
Run BZTyping.py with administrator privileges:
python BZTyping.py- Ctrl+Alt: Start/stop recording
- Ctrl+Shift: Show/hide the UI
- Status Indicator: Shows whether the application is recording, processing, or ready
- Timer: Displays recording duration
- Last Transcription: Shows the most recent transcribed text
- Minimize Button: Switch to minimalist mode
- Settings Button: Configure application preferences
- Close Button: Exit the application
Access the settings dialog to customize:
- Hotkeys: Change the key combinations for recording and UI toggling
- Language: Select from multiple languages for transcription
BZTyping uses the Whisper large language model to convert speech to text. When you press the recording hotkey:
- Your microphone input is captured
- The audio is processed by the Whisper model
- The transcribed text is typed at your current cursor position
- The text also appears in the application interface
- No Transcription: Ensure your microphone is properly connected and selected as the default input device
- Slow Performance: Consider using a smaller model size or ensure GPU acceleration is working
- Hotkeys Not Working: Run the application with administrator privileges
- PyAudio Installation Issues: Follow the manual installation steps provided when the error occurs
This project is open source and available under the MIT License.
- OpenAI Whisper - Speech recognition model
- Faster Whisper - Optimized Whisper implementation
- PyQt5 - GUI framework
- PyAudio - Audio processing library
- Keyboard - Keyboard hook library