A minimal systray dictation tool for Windows (Linux support planned later).
Press a global hotkey to record your voice. The transcribed text is automatically copied to your clipboard and optionally pasted into the active application.
- Faster-Whisper Server running on
localhost:8000 - Python 3.10+
Download the latest MySimpleDictation.exe from the
Releases
page and run it directly -- no installation required, no admin rights needed.
| Action | Result |
|---|---|
| Click tray icon | Show window |
| Right-click tray icon | Show / Quit |
| Press hotkey | Start recording |
| Press hotkey again | Stop & transcribe |
| Close (X) | Quit app (or minimize, see Settings) |
Press the Language Picker hotkey (default: WIN+PGUP) at any time — even when the
app is minimized to tray — to open a small always-on-top popup for quick language
switching.
| Action | Result |
|---|---|
| Press shortcut key (D, E, F, ...) | Select language instantly |
| Mouse click on entry | Select language |
| Arrow keys + Enter | Navigate and select |
| No action for 3s | Popup closes, language unchanged |
Available shortcut keys: A Auto, D Deutsch, E English, F Francais,
I Italiano, S Espanol, P Portugues, N Nederlands. Languages without a key (Polski,
Русский, etc.) can be selected with mouse or arrow keys.
The Language Picker hotkey can be changed in Settings (same record-and-press logic as the main recording hotkey).
The language dropdown in the main window sets the input language -- the language you are speaking. It does NOT set the output language.
- Set it to the language you speak for best accuracy
- Leave it on
Autoto let Whisper detect the language automatically - If you speak English but set it to
de, Whisper will attempt to translate your English into German -- this is not a bug, it is how thelanguageparameter works in Faster-Whisper
Every transcription is saved to transcription.txt next to the executable.
The last transcript is automatically loaded on startup -- close and reopen anytime,
your text will still be there.
Click the gear button to configure:
| Setting | Default | Description |
|---|---|---|
| Whisper URL | http://localhost:8000 |
Faster-Whisper server address |
| Whisper Model | Systran/faster-whisper-small |
Model to use for transcription |
| Hotkey | WIN+PGDN |
Global recording shortcut - click Record and press your combo |
| Hold to record | off | Hold key = record, release = stop |
| Auto-Paste | off | Simulate Ctrl+V after transcription |
| Append newline after paste | off | Send a newline after auto-paste (SHIFT+ENTER / CRLF / ENTER) |
| Close window = minimize | off | X button minimizes to tray instead of quitting |
| Auto-minimize | off | Hide to tray when window loses focus |
| Start minimized | off | Start directly in tray on launch |
| Always on top | off | Keep window above all other windows |
| Icon colors | green / red / yellow | Tray icon colors for ready / recording / blink state |
When Auto-Paste is active, you can optionally send a newline after the pasted text. Three modes are available:
| Mode | Behavior |
|---|---|
SHIFT+ENTER |
New line without submitting (works in Claude, ChatGPT, etc.) |
CRLF |
Sends \r\n directly as text |
ENTER (hard) |
Standard Enter key -- submits forms/chats |
This setting is only available when Auto-Paste is enabled. Disabling Auto-Paste automatically resets this option to off / SHIFT+ENTER.
Language parameter is input language, not output language.
The language selector tells Whisper which language you are speaking. If you set
de (Deutsch) and speak English, Whisper will try to translate your English into
German. This is the intended behavior of the Faster-Whisper language parameter.
Use Auto if you switch between languages.
Translation Trick: This behavior can be used intentionally -- speak German and set the language to
Englishto get a rough on-the-fly translation. Quality depends on the Whisper model. Not a replacement for a dedicated translation tool, but useful for quick results.
| Color | State |
|---|---|
| Green (default #00AF3A) | Ready |
| Red (default #CC0000) | Recording (alternates with blink color) |
| Yellow (default #FFC800) | Recording blink (alternates with red every 500ms) |
Colors are fully customizable in Settings.
| Version | Changes |
|---|---|
| v2.5 | Health check before transcription: warns if Whisper server is unreachable |
| v2.4 | Language Picker popup (WIN+PGUP): global hotkey opens an always-on-top mini window to switch language via keypress, mouse click or arrow+Enter; auto-closes after 3s with no selection; configurable hotkey in Settings |
| v2.3 | Second instance brings first to foreground instead of silently exiting |
| v2.3 | Removed unnecessary, buggy features and fixed pasting issue when using HOLD keyboard mode for better reliability |
| v2.2 | Microphone icon (EXE, taskbar, window, systray), recording blink animation (red/yellow), configurable blink color |
| v2.1 | Mouse dual-button trigger with activation delay, append newline after paste (SHIFT+ENTER / CRLF / ENTER), hold-mode paste delay fix, settings window follows main window (WIN+D / Alt+Tab) |
| v2.0 | Append mode, auto-save, load last transcript on startup, always-on-top |
| v1.8 | Language selector |
| v1.6 | Whisper model selector |
| v1.5 | Start minimized option |
| v1.4 | Custom hotkey, auto-minimize, custom tray icon colors |
| v1.3 | Systray fixes |
| v1.1 | Cross-platform hotkey support |
| v1.0 | Initial release |