Settings

Open Settings from the gear icon at the bottom of the sidebar. It has eight tabs: General, Shortcuts, Audio, Dictation, Privacy & data, Advanced, About, and Debug (hidden by default). Feature settings, like the assistant’s screen vision or AI cleanup’s writing style, live on their sidebar pages, not here. Settings is for preferences you set once and forget.

General

Appearance

Setting	Default	What it does
Appearance	`Light`	`Light`, `Dark`, or `System`.
Text size	`Default`	Scales the whole main window. `Small`, `Default`, `Large`, or `Extra large`. Applied as a webview zoom factor (0.9, 1.0, 1.1, 1.2).
App language	Your system locale	The language of the SpeakoFlow interface. Twenty languages available.

Startup

Setting	Default	What it does
Launch on startup	Off	Open SpeakoFlow when you sign in.
Start hidden	Off	Launch straight to the tray without showing the window.
Show tray icon	On	Keep SpeakoFlow in the system tray or menu bar.
Quit when I close the window	Off	Off hides the window to the tray and keeps hotkeys live. On quits the app.

If the tray icon is off, Start hidden shows the window anyway. Without a tray icon there would be no way back in.

Shortcuts

Shortcuts are grouped into Dictation and Assistant.

Dictation

Shortcut	Windows	macOS	Linux
Dictate	`Left Ctrl + Left Win`	`Fn` (globe)	`Ctrl + Space`
Dictate and clean up	`Left Ctrl + Left Win + Shift`	`Fn + Shift`	`Ctrl + Shift + Space`
Cancel	`Esc`	`Esc`	Not shown

The cleanup shortcut appears only while AI cleanup is on and set to its own shortcut. If cleanup runs on every dictation, this row says so and links to the AI cleanup page. Cancel is hidden on Linux. Cancel and the conversation shortcut can be turned off from the shortcut editor. Hold to talk / Tap to toggle is at the bottom of this group. It applies to every recording shortcut. Default: Hold to talk. See Hold or tap.

Assistant

Shortcut	Windows	macOS	Linux
Ask the assistant	`Left Ctrl + Left Alt`	`Fn + Ctrl`	`Ctrl + Alt + Space`
Start or end a conversation	`Left Ctrl + Left Alt + C`	`Fn + Ctrl + C`	`Ctrl + Alt + C`

Both rows appear only while the assistant is on (it is on by default). If you turned it off, a single row links to the Assistant page. Click any shortcut’s keys to record a new one. The editor also offers “Use default” and, for Cancel and the conversation shortcut, “Turn off”. See Keyboard shortcuts.

Audio

Microphone

Setting	Default	What it does
Microphone	System default	Which input device to record from.
Always-on microphone	Off	Keep the audio stream open between recordings for faster back-to-back dictation. May affect Bluetooth audio.
Mute while recording	Off	Mute system audio while you record so playback does not bleed into the dictation.

Sounds

Setting	Default	What it does
Audio feedback	Off	Play a sound when recording starts and stops. The rows below stay disabled until you turn this on.
Sound theme	`Default`	The start and stop cue: `Default`, `Marimba`, `Pop`, or `Click`. A play button previews the pair.
Output device	System default	Which device plays feedback sounds.
Volume	100%	Volume of feedback sounds. This does not affect the assistant’s spoken replies.

Dictation

Overlay

The small pill or card that shows while you record.

Setting	Default	What it does
Overlay style	`Automatic`	`None`, `Minimal` (compact pill), `Live` (card showing your words as you speak), or `Automatic` (Live when the model supports streaming, Minimal otherwise).
Stay on screen	`Quick` (1 s)	How long the finished Live card stays up. `Quick`, `Standard` (3 s), `Long` (6 s), or `Extended` (10 s). Hovering holds it regardless. Only shown when the overlay is Live.
Overlay position	`Bottom` (`None` on Linux)	Where the overlay sits on screen.

See The recording overlay and Live transcription.

Typing and pasting

Setting	Default	What it does
Paste method	`Ctrl+V` (Windows/macOS), `Direct` (Linux)	How the transcript reaches the active app. See Output.
Typing tool	`Auto`	Linux only. Which tool types text into apps: Auto, wtype, kwtype, dotool, ydotool, or xdotool.
Clipboard handling	`Don't modify`	`Don't modify` restores your clipboard after pasting. `Copy to clipboard` leaves the transcript there.
Append trailing space	Off	Add a space after each dictation so the next one does not run into it.
Auto-submit	Off	Press Enter (or another key) after pasting, for chat fields that send on Enter.

Extras

Setting	Default	What it does
Spoken emojis	Off	Turn spoken commands like “happy emoji” into their Unicode emoji.

Generate with Flow

At the bottom of the Dictation tab. Start a dictation with a trigger phrase (default “Hey Flow”) and the finished text is pasted instead of your words. Off by default. See Generate with Flow.

Privacy & data

Recording retention

Controls how long unstarred dictation recordings are kept.

Setting	Default	What it does
Retention policy	`Keep a set number`	`Keep a set number`, `Never delete`, `After 3 days`, `After 2 weeks`, `After 3 months`, or a custom number of days.
Recording limit	20	How many unstarred recordings to keep (1 to 1000). Only shown for “Keep a set number”.

Starred recordings are always kept. Changes apply right away, and the UI shows how many recordings will be deleted before you confirm. Recordings folder opens the folder where recordings are stored on disk.

Folders

Row	What it shows
App data folder	Where settings, models, and history live. A button opens it.
Log folder	Where the log file is written. A button opens it.

Advanced

Performance

Setting	Default	What it does
Model unload timeout	`After 5 minutes`	How long idle before the local LLM engine (assistant and cleanup) frees its memory. `Never`, `Immediately`, 2/5/10/15 minutes, or 1 hour. See Acceleration & GPU.
Experimental features	Off	Reveals the Experimental group below.

Experimental

This group appears only when Experimental features is on.

Setting	Default	What it does
Keyboard engine	`Tauri` on Linux, `SpeakoFlow Keys` elsewhere	Which backend listens for global shortcuts. See Troubleshooting.
Acceleration	`Auto`	Which compute device the speech model runs on. See Acceleration & GPU.
Keep mic open between transcriptions	Off	Leaves the audio stream open after recording. Can improve back-to-back dictation speed.

About

Shows the app name, version, and an update status line. If an update is available, you can install or download it here.

Row	What it does
Check for updates automatically	On by default. Checks 8 seconds after launch, then every 12 hours.
Send feedback	Opens the feedback dialog. See About.
Source code	Links to the GitHub repository.
License	Links to the MIT license file.
Documentation	Opens the docs site.

Debug

Hidden by default. Press Ctrl + Shift + D (Cmd + Shift + D on macOS) with the Settings window focused to reveal it. Full details in Debug & advanced.

Accuracy

Models

Help

Settings

General

Appearance

Startup

Shortcuts

Dictation

Assistant

Audio

Microphone

Sounds

Dictation

Overlay

Typing and pasting

Extras

Generate with Flow

Privacy & data

Recording retention

Folders

Advanced

Performance

Experimental

About

Debug

​General

​Appearance

​Startup

​Shortcuts

​Dictation

​Assistant

​Audio

​Microphone

​Sounds

​Dictation

​Overlay

​Typing and pasting

​Extras

​Generate with Flow

​Privacy & data

​Recording retention

​Folders

​Advanced

​Performance

​Experimental

​About

​Debug

General

Appearance

Startup

Shortcuts

Dictation

Assistant

Audio

Microphone

Sounds

Dictation

Overlay

Typing and pasting

Extras

Generate with Flow

Privacy & data

Recording retention

Folders

Advanced

Performance

Experimental

About

Debug