Dictation Tray (Windows)

A small Windows tray application for speech-to-text dictation.

The application records microphone audio on demand, transcribes it locally using an AI model, and copies the resulting text to the clipboard.

No API keys or cloud services are required.

---

Just a note: AI wrote this readme. It seems fine]

I just wanted to build a quick speech-to-text utility for myself and ChatGPT helped make this python script. It lives in your system tray and simply lets you dictate and outputs text.

I've been using it all day and it's working very well.

---

Requirements

• Windows 10 or Windows 11
• Python 3.10 or newer
• A working microphone

---

Installation

1. Clone the repository

git clone https://github.com/YOUR_USERNAME/dictation-tray.git
cd dictation-tray

2. Install Python dependencies

pip install pyside6 sounddevice numpy pyperclip faster-whisper

3. Install FFmpeg

FFmpeg is required for audio processing.

Using winget:

winget install Gyan.FFmpeg

Restart your terminal after installing FFmpeg.

---

Running the Application

python dictate_tray.py

• A window will open
• A system tray icon will appear
• Closing the window hides it to the tray instead of exiting

---

Usage

1. Click Start
2. Speak into the microphone
3. Click Stop
4. The transcribed text appears in the window and is copied to the clipboard

Paste the text anywhere using Ctrl + V.

---

Configuration

Configuration is defined in the Config class at the top of dictate_tray.py.

class Config:
    model_size = "small"
    device = "cpu"
    compute_type = "int8"
    language = "en"

Options

• model_size: tiny, base, small, medium, large-v3
• device: cpu or cuda
• language: set to None for auto-detection

---

Notes

• Transcription is performed locally
• Audio is not uploaded or stored
• The application only records when explicitly started

---

License

MIT License