diff --git a/src/calibre/gui2/tts/linux_config.py b/src/calibre/gui2/tts/linux_config.py
index 1b69f8d2bc..4d5e3e8429 100644
--- a/src/calibre/gui2/tts/linux_config.py
+++ b/src/calibre/gui2/tts/linux_config.py
@@ -18,7 +18,10 @@ class VoicesModel(QAbstractTableModel):
     def __init__(self, voice_data, default_output_module, parent=None):
         super().__init__(parent)
         self.voice_data = voice_data
-        self.current_voices = voice_data[default_output_module]
+        try:
+            self.current_voices = voice_data[default_output_module]
+        except KeyError as e:
+            raise ValueError(_('Speech dispatcher on this system is not configured with any available voices. Install some voices first.')) from e
         self.column_headers = (_('Name'), _('Language'), _('Variant'))
 
     def rowCount(self, parent=None):