Skip to content

Troubleshooting

Known Limitations

Maximum recording length is 30 seconds

Speed of Sound currently caps each recording at 30 seconds. When that limit is reached, transcription is triggered automatically, the same as if you had stopped recording manually. This is not configurable. For most voice-typing use cases this is plenty, but if longer recordings matter to you, please 👍 this issue to help prioritize it.

Only the default microphone is supported

Speed of Sound records from your system's default microphone. There currently is no in-app microphone selector. To use a different input device, change the default in your system sound settings before dictating.

Common Issues

Remote desktop portal is not supported

Symptom: A banner at the top of the main window says "Remote desktop portal is not supported."

Why this happens: Speed of Sound relies on the XDG Remote Desktop Portal standard to type text into other applications. This portal must be supported by your desktop environment's portal backend. However, not all desktop environments ship a backend that implements it (this table provides a good compatibility matrix).

How to fix it: We recommend reporting the missing support to your desktop environment's issue tracker. In the meantime, if possible, consider switching to a desktop environment that implements this portal (e.g. GNOME, KDE). We are also exploring alternatives such as clipboard-based text input. If that would be useful to you, let us know.

Non-Latin text produces only spaces and punctuation

Symptom: Dictating in a non-Latin language (e.g., Cyrillic, Arabic, CJK) outputs only spaces and punctuation. Latin-script languages work fine.

Why this happens: Speed of Sound sends keystrokes through the XDG Remote Desktop Portal. The compositor interprets them using the currently active keyboard layout. If no matching layout is active, non-Latin keysyms are silently dropped.

How to fix it: Add the appropriate keyboard layout in your system's keyboard or input source settings and switch to it before dictating. For example, on GNOME, this is under Settings → Keyboard → Input Sources. You can switch between input sources with Super + Space.

Transcribed text is incomplete

Symptom: Typically, the beginning of the text is missing.

Why this happens: After transcription, Speed of Sound minimizes or hides its main window, waits briefly to give the desktop environment time to do so, and then starts typing. If the wait is too short, the first few characters will be lost because they land in the Speed of Sound window rather than your target application.

How to fix it: Increase the Post-hide delay in Preferences → Advanced. This is the pause between the window hiding and the first keystroke. Alternatively, enable Record in background (see above).

Some characters are missing or appear out of order

Symptom: The transcribed text arrives with letters dropped or jumbled.

Why this happens: Some applications cannot process keystrokes as fast as Speed of Sound sends them, causing characters to be lost or reordered.

How to fix it: Increase the Typing delay in Preferences → Advanced. This adds a pause between each individual keystroke. You can also increase this value if you simply like the effect of text appearing more gradually.

Speed of Sound appears on the wrong workspace and doesn't type into my active application

Symptom: On a multi-workspace setup, the main Speed of Sound window opens on a different workspace, and transcribed text lands in the wrong application or nowhere at all.

Why this happens: When a minimized window is restored, some desktop environments bring it back on its original workspace rather than the current one, stealing focus and breaking text input. By default, Speed of Sound minimizes the main window before typing.

How to fix it: Enable Hide instead of minimize in Preferences → General → App Behavior. A hidden window is typically restored on the active workspace instead.

Trade-off: The app no longer appears in the dock. To bring the main window back, you can, for example, use a global shortcut (Preferences → General → Global Shortcut).

I cannot change or reset my global shortcut

Symptom: A global shortcut was set up via the Preferences dialog, but the shortcut no longer works (or was set incorrectly), and there is no way to change or clear it from the UI.

Why this happens: Speed of Sound uses the ConfigureShortcuts portal method to let you reassign your shortcut after initial setup. This method was added recently and is only available on newer desktop environments, so Speed of Sound only shows the Configure button when it detects support for it. If the button is absent, your desktop does not yet expose that capability.

How to fix it: You have two options:

  1. Switch to the manual method. Instead of the automatic portal-based setup, follow Option 2 in the keyboard shortcut documentation to assign a shortcut directly in your desktop environment settings. This gives you full control over the key combination and lets you change or remove it at any time.

  2. Reset the shortcut from the command line. Follow the testing instructions from the Stargate project, replacing the application ID with io.speedofsound.SpeedOfSound. This will delete the existing shortcut binding. Restart Speed of Sound afterward, the Set Up button will reappear, and you can configure a new shortcut.

I don't want the Speed of Sound main window to show every time I dictate.

Symptom: The main window appears on every dictation, which feels intrusive for frequent voice typing.

Why this happens: By default, Speed of Sound presents its main window each time you trigger dictation to show recording and transcription progress.

How to fix it: Enable Record in background in Preferences → General → App Behavior. When active, recording, transcription, and typing all run silently without showing the main window.

Note: Because Speed of Sound uses the Desktop Portals standard, your desktop environment will still display a microphone in-use indicator while recording is active and while the app has typing permissions. This serves as a lightweight substitute for the in-app progress display.