fix(stt): flag empty transcripts as no_speech

An STT provider that hears no words is reporting silence, not failing.
ElevenLabs/xAI empty-transcript errors now carry no_speech so live
voice loops can re-listen quietly instead of surfacing an error on
every pause.

tests/tools/test_transcription_tools.py

@@ -1209,6 +1209,7 @@ class TestTranscribeXAI:
 
         assert result["success"] is False
         assert "empty transcript" in result["error"]
+        assert result["no_speech"] is True  # live voice loops treat this as silence
 
     def test_permission_error(self, monkeypatch, sample_ogg, mock_xai_http_module):
         monkeypatch.setenv("XAI_API_KEY", "xai-test-key")
@@ -1453,6 +1454,7 @@ class TestTranscribeElevenLabs:
 
         assert result["success"] is False
         assert "empty transcript" in result["error"]
+        assert result["no_speech"] is True  # live voice loops treat this as silence
 
 
 # ============================================================================

tools/transcription_tools.py

@@ -1546,6 +1546,7 @@ def _transcribe_xai(file_path: str, model_name: str) -> Dict[str, Any]:
                 "success": False,
                 "transcript": "",
                 "error": "xAI STT returned empty transcript",
+                "no_speech": True,
             }
 
         logger.info(
@@ -1633,6 +1634,7 @@ def _transcribe_elevenlabs(file_path: str, model_name: str) -> Dict[str, Any]:
                 "success": False,
                 "transcript": "",
                 "error": "ElevenLabs STT returned empty transcript",
+                "no_speech": True,
             }
 
         logger.info(