livekit-plugins-google 0.11.0__py3-none-any.whl → 0.11.2__py3-none-any.whl

livekit/plugins/google/beta/realtime/api_proto.py

@@ -6,7 +6,7 @@ from google.genai import types
 
 from ..._utils import _build_gemini_ctx, _build_tools
 
-LiveAPIModels = Literal["gemini-2.0-flash-001",]
+LiveAPIModels = Literal["gemini-2.0-flash-exp"]
 
 Voice = Literal["Puck", "Charon", "Kore", "Fenrir", "Aoede"]
 

livekit/plugins/google/beta/realtime/realtime_api.py

@@ -83,6 +83,7 @@ class Capabilities:
 class ModelOptions:
     model: LiveAPIModels | str
     api_key: str | None
+    api_version: str
     voice: Voice | str
     response_modalities: list[Modality] | None
     vertexai: bool
@@ -107,6 +108,7 @@ class RealtimeModel:
         instructions: str | None = None,
         model: LiveAPIModels | str = "gemini-2.0-flash-exp",
         api_key: str | None = None,
+        api_version: str = "v1alpha",
         voice: Voice | str = "Puck",
         modalities: list[Modality] = [Modality.AUDIO],
         enable_user_audio_transcription: bool = True,
@@ -136,6 +138,7 @@ class RealtimeModel:
         Args:
             instructions (str, optional): Initial system instructions for the model. Defaults to "".
             api_key (str or None, optional): Google Gemini API key. If None, will attempt to read from the environment variable GOOGLE_API_KEY.
+            api_version (str, optional): The version of the API to use. Defaults to "v1alpha".
             modalities (list[Modality], optional): Modalities to use, such as ["TEXT", "AUDIO"]. Defaults to ["AUDIO"].
             model (str or None, optional): The name of the model to use. Defaults to "gemini-2.0-flash-exp".
             voice (api_proto.Voice, optional): Voice setting for audio outputs. Defaults to "Puck".
@@ -187,6 +190,7 @@ class RealtimeModel:
         self._rt_sessions: list[GeminiRealtimeSession] = []
         self._opts = ModelOptions(
             model=model,
+            api_version=api_version,
             api_key=self._api_key,
             voice=voice,
             enable_user_audio_transcription=enable_user_audio_transcription,
@@ -289,7 +293,7 @@ class GeminiRealtimeSession(utils.EventEmitter[EventTypes]):
             tools=tools,
         )
         self._client = genai.Client(
-            http_options=HttpOptions(api_version="v1alpha"),
+            http_options=HttpOptions(api_version=self._opts.api_version),
             api_key=self._opts.api_key,
             vertexai=self._opts.vertexai,
             project=self._opts.project,

livekit/plugins/google/stt.py

@@ -94,7 +94,7 @@ class STT(stt.STT):
         punctuate: bool = True,
         spoken_punctuation: bool = False,
         model: SpeechModels | str = "latest_long",
-        location: str = "us-central1",
+        location: str = "global",
         sample_rate: int = 16000,
         credentials_info: dict | None = None,
         credentials_file: str | None = None,
@@ -114,7 +114,7 @@ class STT(stt.STT):
             punctuate(bool): whether to punctuate the audio (default: True)
             spoken_punctuation(bool): whether to use spoken punctuation (default: False)
             model(SpeechModels): the model to use for recognition default: "latest_long"
-            location(str): the location to use for recognition default: "us-central1"
+            location(str): the location to use for recognition default: "global"
             sample_rate(int): the sample rate of the audio default: 16000
             credentials_info(dict): the credentials info to use for recognition (default: None)
             credentials_file(str): the credentials file to use for recognition (default: None)
@@ -488,7 +488,6 @@ class SpeechStream(stt.SpeechStream):
                             ),
                         ),
                         streaming_features=cloud_speech.StreamingRecognitionFeatures(
-                            enable_voice_activity_events=True,
                             interim_results=self._config.interim_results,
                         ),
                     )

livekit/plugins/google/version.py

@@ -12,4 +12,4 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
-__version__ = "0.11.0"
+__version__ = "0.11.2"

{livekit_plugins_google-0.11.0.dist-info → livekit_plugins_google-0.11.2.dist-info}/METADATA

@@ -1,6 +1,6 @@
-Metadata-Version: 2.2
+Metadata-Version: 2.4
 Name: livekit-plugins-google
-Version: 0.11.0
+Version: 0.11.2
 Summary: Agent Framework plugin for services from Google Cloud
 Home-page: https://github.com/livekit/agents
 License: Apache-2.0

{livekit_plugins_google-0.11.0.dist-info → livekit_plugins_google-0.11.2.dist-info}/RECORD

@@ -4,15 +4,15 @@ livekit/plugins/google/llm.py,sha256=LZaHsrkjfboRZLWm7L2G0mw62q2sXBNj4YeeV2Sk2uU
 livekit/plugins/google/log.py,sha256=GI3YWN5YzrafnUccljzPRS_ZALkMNk1i21IRnTl2vNA,69
 livekit/plugins/google/models.py,sha256=SGjAumdDK97NNLwMFcqZdKR68f1NoGB2Rk1UP2-imG0,1457
 livekit/plugins/google/py.typed,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
-livekit/plugins/google/stt.py,sha256=96GJmGDAIBdCpDECArwIXpj2s1xlcA_zuvTnwsvq4xA,22854
+livekit/plugins/google/stt.py,sha256=l6UB9oaM7eFInnI_0t7Ub-edXLVRgvaiyHj-e_gEuwE,22781
 livekit/plugins/google/tts.py,sha256=pG9_pibO3NDGEMa4huU5S9lbeyI3daQyrS17SuTKfZI,8008
-livekit/plugins/google/version.py,sha256=BvmVdoHkxksDSQP-uWrqIiyaAUImEyxSohntkIBNZRo,601
+livekit/plugins/google/version.py,sha256=_06ctkD1XWTWec2BVgcsxun2sFLxqnvJJJs7ZxIBuHA,601
 livekit/plugins/google/beta/__init__.py,sha256=AxRYc7NGG62Tv1MmcZVCDHNvlhbC86hM-_yP01Qb28k,47
 livekit/plugins/google/beta/realtime/__init__.py,sha256=sGTn6JFNyA30QUXBZ_BV3l2eHpGAzR35ByXxg77vWNU,205
-livekit/plugins/google/beta/realtime/api_proto.py,sha256=ralrRZqIbE71oyuLKRYaXHvm6tcHMwBJueKvSO8Xfus,658
-livekit/plugins/google/beta/realtime/realtime_api.py,sha256=vZHiWNk8PorxtrHSmA7Ya6ZvCjT37YSJN-MxK8ebdrs,22795
+livekit/plugins/google/beta/realtime/api_proto.py,sha256=9EhmwgeIgKDqdSijv5Q9pgx7UhAakK02ZDwbnUsra_o,657
+livekit/plugins/google/beta/realtime/realtime_api.py,sha256=8JdWUMUheGhy1ia6JbN3_U2_cL7CNs8-1fTOAgW4I38,22999
 livekit/plugins/google/beta/realtime/transcriber.py,sha256=rjXO0cSPr3HATxrSfv1MX7IbrjmiTvnLPF280BfRBL8,9809
-livekit_plugins_google-0.11.0.dist-info/METADATA,sha256=b8Aj_eQnGhAT3DQa77KLHZBDGAWZYdrnTBWjVODAm2k,3732
-livekit_plugins_google-0.11.0.dist-info/WHEEL,sha256=jB7zZ3N9hIM9adW7qlTAyycLYW9npaWKLRzaoVcLKcM,91
-livekit_plugins_google-0.11.0.dist-info/top_level.txt,sha256=OoDok3xUmXbZRvOrfvvXB-Juu4DX79dlq188E19YHoo,8
-livekit_plugins_google-0.11.0.dist-info/RECORD,,
+livekit_plugins_google-0.11.2.dist-info/METADATA,sha256=MQF9voerbBB1t5fGRw94z7jyfgJOnsM-DmWxtCT10V8,3732
+livekit_plugins_google-0.11.2.dist-info/WHEEL,sha256=1tXe9gY0PYatrMPMDd6jXqjfpz_B-Wqm32CPfRC58XU,91
+livekit_plugins_google-0.11.2.dist-info/top_level.txt,sha256=OoDok3xUmXbZRvOrfvvXB-Juu4DX79dlq188E19YHoo,8
+livekit_plugins_google-0.11.2.dist-info/RECORD,,

{livekit_plugins_google-0.11.0.dist-info → livekit_plugins_google-0.11.2.dist-info}/WHEEL

@@ -1,5 +1,5 @@
 Wheel-Version: 1.0
-Generator: setuptools (75.8.2)
+Generator: setuptools (77.0.3)
 Root-Is-Purelib: true
 Tag: py3-none-any