Doctor Voice Notes

Doctor Voice Notes is an offline Windows desktop application that converts spoken English into text in real time and continuously saves the transcription into a Microsoft Word document.

Designed specifically for healthcare professionals, the application provides a simple, distraction-free interface that enables doctors to dictate patient notes naturally while keeping all data on the local computer.

Features

• Real-time offline speech recognition
• Live transcription display
• Automatic Word document generation
• Autosave every few seconds
• Voice commands for formatting and editing
• Large, easy-to-use interface
• Configurable microphone and save location
• Local settings storage
• Logging and error recovery
• No internet connection required
• No cloud services
• No external APIs
• No user accounts

Technology Stack

• Python 3.12
• Tkinter
• faster-whisper
• sounddevice
• python-docx
• ctranslate2
• PyInstaller

Project Structure

DoctorVoiceNotes/
├── main.py # App entry point
├── build.spec # PyInstaller build spec, onedir mode
├── build.bat # Windows build script
├── installer.iss # Inno Setup installer script
├── requirements.txt # Pinned dependencies
├── README.md # Developer setup, build, and known-issues doc
│
├── assets/
│ └── icons/
│ └── app_icon.ico
│
├── models/
│ └── small.en/ # Speech model
│ ├── model.bin
│ ├── config.json
│ ├── tokenizer.json
│ ├── vocabulary.txt
│ └── README_DOWNLOAD_MODEL.txt
│
├── src/
│ ├── init.py
│ ├── audio/
│ │ ├── init.py
│ │ └── recorder.py # Microphone detection + capture
│ ├── transcription/
│ │ ├── init.py
│ │ └── engine.py # Whisper transcription + VAD
│ ├── document/
│ │ ├── init.py
│ │ └── writer.py # .docx writing + autosave
│ ├── commands/
│ │ ├── init.py
│ │ └── voice_commands.py # Voice command parsing
│ ├── ui/
│ │ ├── init.py
│ │ ├── main_window.py # Main application window
│ │ └── settings_dialog.py # Settings dialog
│ └── utils/
│ ├── init.py
│ ├── config_manager.py # settings.json load/save
│ ├── logger.py # Rotating file logging
│ └── paths.py # Frozen-vs-source path resolution
│
├── config/ # Created automatically on first run
│ └── settings.json
├── logs/ # Created automatically on first run
│ └── app.log
├── notes/ # Created automatically on first run
│ └── PatientNotes.docx
│
├── venv/ # Local virtual environment (gitignored)
├── build/ # PyInstaller intermediate output (gitignored)
├── dist/ # PyInstaller final output (gitignored)
│ └── DoctorVoiceNotes/
│ ├── DoctorVoiceNotes.exe
│ ├── _internal/ # Bundled Python + dependencies
│ └── models/small.en/ # Model, copied here by build.bat
└── installer_output/ # Compiled installer (gitignored)
└── DoctorVoiceNotes_Setup_v1.0.0.exe

Not committed to source control (see .gitignore): venv/, build/, dist/, installer_output/, models/small.en/* (except README_DOWNLOAD_MODEL.txt), and the auto-generated config/, logs/, notes/ contents — these are either environment-specific, build artifacts, or the doctor's own clinical data.

Installation

1. Download the latest release.
2. Run the DoctorVoiceNotes.exe file.
3. If Windows SmartScreen appears, click More info → Run anyway.
4. Wait for the application to load the speech recognition model.
5. Click Start to begin dictating.

No Python installation, internet connection, or additional setup is required.

Configuration

Application settings are stored locally and include:

• Default save folder
• Document name
• Microphone selection
• Autosave interval
• Theme
• Font size
• Voice command settings

Voice Commands

Supported commands include:

• New paragraph
• Full stop
• Comma
• New patient
• Date
• Time
• Save note
• Delete last sentence

Privacy

Doctor Voice Notes is designed with privacy as a primary requirement.

• Fully offline operation
• No cloud storage
• No telemetry
• No internet communication
• No patient data leaves the local computer

Roadmap

Planned improvements include:

• Patient-specific documents
• PDF export
• Medical abbreviation expansion
• Prescription mode
• Foot pedal support
• Local encrypted backups
• Offline medical terminology dictionary

License

This project is proprietary software developed for a private client.