livekit-plugins-volcenginee 1.3.0__tar.gz → 1.3.1__tar.gz

{livekit_plugins_volcenginee-1.3.0 → livekit_plugins_volcenginee-1.3.1}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: livekit-plugins-volcenginee
-Version: 1.3.0
+Version: 1.3.1
 Summary: LiveKit Agent Plugins for Volcengine
 Author-email: linsanzhu <2287225730@qq.com>
 Keywords: audio,livekit,realtime,video,webrtc
@@ -38,6 +38,8 @@ Description-Content-Type: text/markdown
 - `volcengine.TTS` 默认音色为 `zh_female_xiaohe_uranus_bigtts`
 - `volcengine.STT` 现已直接使用豆包大模型流式语音识别接口
 - `volcengine.STT` 已补齐新版握手头 `X-Api-Connect-Id`
+- `volcengine.STT` 切换到新版控制台鉴权，仅使用 `X-Api-Key` 头
+- `volcengine.TTS` 同步切换到新版控制台鉴权，仅使用 `X-Api-Key` 头
 
 ## ✨ 特性
 
@@ -95,8 +97,8 @@ pip install -e .
 
 | 环境变量 | 描述 | 获取方式 |
 |----------|------|----------|
-| `VOLCENGINE_TTS_ACCESS_TOKEN` | TTS 服务的访问令牌 | [语音合成控制台](https://console.volcengine.com/speech/service/16) |
-| `VOLCENGINE_STT_ACCESS_TOKEN` | STT 服务的访问令牌 | [语音识别控制台](https://console.volcengine.com/speech/service/16) |
+| `VOLCENGINE_TTS_API_KEY` | TTS 服务的 API Key（新版控制台） | [语音合成控制台](https://console.volcengine.com/speech/service/16) |
+| `VOLCENGINE_STT_API_KEY` | STT 服务的 API Key（新版控制台） | [语音识别控制台](https://console.volcengine.com/speech/service/16) |
 | `VOLCENGINE_LLM_API_KEY` | LLM 服务的 API 密钥 | [大模型控制台](https://console.volcengine.com/ark/) |
 | `VOLCENGINE_REALTIME_ACCESS_TOKEN` | 实时服务的访问令牌 | [实时语音控制台](https://console.volcengine.com/speech/service/10011) |
 | `VOLCENGINE_TTS_RESOURCE_ID` | TTS 资源 ID，默认 `seed-tts-2.0` | [豆包语音合成接口文档](https://www.volcengine.com/docs/6561/1257584) |
@@ -105,8 +107,8 @@ pip install -e .
 
 ```bash
 # .env
-VOLCENGINE_TTS_ACCESS_TOKEN=your_tts_token_here
-VOLCENGINE_STT_ACCESS_TOKEN=your_stt_token_here
+VOLCENGINE_TTS_API_KEY=your_tts_api_key_here
+VOLCENGINE_STT_API_KEY=your_stt_api_key_here
 VOLCENGINE_LLM_API_KEY=your_llm_api_key_here
 VOLCENGINE_REALTIME_ACCESS_TOKEN=your_realtime_token_here
 ```
@@ -152,11 +154,11 @@ async def entry_point(ctx: JobContext):
     agent = Agent(instructions="You are a helpful assistant.")
 
     # 使用豆包大模型流式语音识别
-    stt = volcengine.STT(app_id="your_app_id")
+    stt = volcengine.STT(api_key="your_api_key")
 
     session = AgentSession(
         stt=stt,
-        tts=volcengine.TTS(app_id="your_tts_app_id"),
+        tts=volcengine.TTS(api_key="your_tts_api_key"),
         llm=volcengine.LLM(model="doubao-1-5-pro-32k-250115")
     )
 
@@ -366,12 +368,12 @@ async def entry_point(ctx: JobContext):
     session = AgentSession(
         # 语音识别
         stt=volcengine.STT(
-            app_id="your_stt_app_id",
+            api_key="your_stt_api_key",
             resource_id="volc.seedasr.sauc.duration"
         ),
         # 语音合成
         tts=volcengine.TTS(
-            app_id="your_tts_app_id",
+            api_key="your_tts_api_key",
             resource_id="seed-tts-2.0",
             voice="zh_female_xiaohe_uranus_bigtts"
         ),
@@ -395,9 +397,8 @@ if __name__ == "__main__":
 
 ```python
 volcengine.TTS(
-    app_id: str,           # 应用ID
-    access_token: str | None = None,    # 访问令牌，可用环境变量
-    resource_id: str | None = None,     # 默认 seed-tts-2.0
+    api_key: str,                    # API Key（新版控制台）
+    resource_id: str | None = None,  # 默认 seed-tts-2.0
     voice: str = "zh_female_xiaohe_uranus_bigtts",
     speed: float = 1.0,
     volume: float = 1.0,
@@ -421,8 +422,7 @@ volcengine.TTS(
 
 ```python
 volcengine.STT(
-    app_id: str | None = None,
-    access_token: str | None = None,
+    api_key: str | None = None,        # API Key（新版控制台），可用环境变量 VOLCENGINE_STT_API_KEY
     resource_id: str | None = None,  # 例如 volc.bigasr.sauc.duration / volc.seedasr.sauc.duration
     model_name: str = "bigmodel",
     enable_itn: bool = False,
@@ -588,6 +588,8 @@ A: 根据您的应用需求选择合适的版本：
 - 修复无音频返回时静默失败的问题
 - 删除旧版普通 STT，实现统一为豆包大模型流式语音识别
 - `volcengine.STT` 补齐 `X-Api-Connect-Id` 握手头
+- `volcengine.STT` 切换到新版控制台鉴权，移除 `app_id` / `access_token`，改用单一 `api_key` 并发送 `X-Api-Key` 头
+- `volcengine.TTS` 同步切换到新版控制台鉴权，移除 `app_id` / `access_token`，改用单一 `api_key` 并发送 `X-Api-Key` 头
 
 ## 🤝 贡献
 

{livekit_plugins_volcenginee-1.3.0 → livekit_plugins_volcenginee-1.3.1}/README.md

@@ -16,6 +16,8 @@
 - `volcengine.TTS` 默认音色为 `zh_female_xiaohe_uranus_bigtts`
 - `volcengine.STT` 现已直接使用豆包大模型流式语音识别接口
 - `volcengine.STT` 已补齐新版握手头 `X-Api-Connect-Id`
+- `volcengine.STT` 切换到新版控制台鉴权，仅使用 `X-Api-Key` 头
+- `volcengine.TTS` 同步切换到新版控制台鉴权，仅使用 `X-Api-Key` 头
 
 ## ✨ 特性
 
@@ -73,8 +75,8 @@ pip install -e .
 
 | 环境变量 | 描述 | 获取方式 |
 |----------|------|----------|
-| `VOLCENGINE_TTS_ACCESS_TOKEN` | TTS 服务的访问令牌 | [语音合成控制台](https://console.volcengine.com/speech/service/16) |
-| `VOLCENGINE_STT_ACCESS_TOKEN` | STT 服务的访问令牌 | [语音识别控制台](https://console.volcengine.com/speech/service/16) |
+| `VOLCENGINE_TTS_API_KEY` | TTS 服务的 API Key（新版控制台） | [语音合成控制台](https://console.volcengine.com/speech/service/16) |
+| `VOLCENGINE_STT_API_KEY` | STT 服务的 API Key（新版控制台） | [语音识别控制台](https://console.volcengine.com/speech/service/16) |
 | `VOLCENGINE_LLM_API_KEY` | LLM 服务的 API 密钥 | [大模型控制台](https://console.volcengine.com/ark/) |
 | `VOLCENGINE_REALTIME_ACCESS_TOKEN` | 实时服务的访问令牌 | [实时语音控制台](https://console.volcengine.com/speech/service/10011) |
 | `VOLCENGINE_TTS_RESOURCE_ID` | TTS 资源 ID，默认 `seed-tts-2.0` | [豆包语音合成接口文档](https://www.volcengine.com/docs/6561/1257584) |
@@ -83,8 +85,8 @@ pip install -e .
 
 ```bash
 # .env
-VOLCENGINE_TTS_ACCESS_TOKEN=your_tts_token_here
-VOLCENGINE_STT_ACCESS_TOKEN=your_stt_token_here
+VOLCENGINE_TTS_API_KEY=your_tts_api_key_here
+VOLCENGINE_STT_API_KEY=your_stt_api_key_here
 VOLCENGINE_LLM_API_KEY=your_llm_api_key_here
 VOLCENGINE_REALTIME_ACCESS_TOKEN=your_realtime_token_here
 ```
@@ -130,11 +132,11 @@ async def entry_point(ctx: JobContext):
     agent = Agent(instructions="You are a helpful assistant.")
 
     # 使用豆包大模型流式语音识别
-    stt = volcengine.STT(app_id="your_app_id")
+    stt = volcengine.STT(api_key="your_api_key")
 
     session = AgentSession(
         stt=stt,
-        tts=volcengine.TTS(app_id="your_tts_app_id"),
+        tts=volcengine.TTS(api_key="your_tts_api_key"),
         llm=volcengine.LLM(model="doubao-1-5-pro-32k-250115")
     )
 
@@ -344,12 +346,12 @@ async def entry_point(ctx: JobContext):
     session = AgentSession(
         # 语音识别
         stt=volcengine.STT(
-            app_id="your_stt_app_id",
+            api_key="your_stt_api_key",
             resource_id="volc.seedasr.sauc.duration"
         ),
         # 语音合成
         tts=volcengine.TTS(
-            app_id="your_tts_app_id",
+            api_key="your_tts_api_key",
             resource_id="seed-tts-2.0",
             voice="zh_female_xiaohe_uranus_bigtts"
         ),
@@ -373,9 +375,8 @@ if __name__ == "__main__":
 
 ```python
 volcengine.TTS(
-    app_id: str,           # 应用ID
-    access_token: str | None = None,    # 访问令牌，可用环境变量
-    resource_id: str | None = None,     # 默认 seed-tts-2.0
+    api_key: str,                    # API Key（新版控制台）
+    resource_id: str | None = None,  # 默认 seed-tts-2.0
     voice: str = "zh_female_xiaohe_uranus_bigtts",
     speed: float = 1.0,
     volume: float = 1.0,
@@ -399,8 +400,7 @@ volcengine.TTS(
 
 ```python
 volcengine.STT(
-    app_id: str | None = None,
-    access_token: str | None = None,
+    api_key: str | None = None,        # API Key（新版控制台），可用环境变量 VOLCENGINE_STT_API_KEY
     resource_id: str | None = None,  # 例如 volc.bigasr.sauc.duration / volc.seedasr.sauc.duration
     model_name: str = "bigmodel",
     enable_itn: bool = False,
@@ -566,6 +566,8 @@ A: 根据您的应用需求选择合适的版本：
 - 修复无音频返回时静默失败的问题
 - 删除旧版普通 STT，实现统一为豆包大模型流式语音识别
 - `volcengine.STT` 补齐 `X-Api-Connect-Id` 握手头
+- `volcengine.STT` 切换到新版控制台鉴权，移除 `app_id` / `access_token`，改用单一 `api_key` 并发送 `X-Api-Key` 头
+- `volcengine.TTS` 同步切换到新版控制台鉴权，移除 `app_id` / `access_token`，改用单一 `api_key` 并发送 `X-Api-Key` 头
 
 ## 🤝 贡献
 

{livekit_plugins_volcenginee-1.3.0 → livekit_plugins_volcenginee-1.3.1}/livekit/plugins/volcengine/stt.py

@@ -76,8 +76,7 @@ def generate_before_payload(sequence: int):
 
 @dataclass
 class STTOptions:
-    app_id: str | None = None
-    access_token: str | None = None
+    api_key: str | None = None
     source_type: Literal["duration", "concurrent"] = "duration"
     resource_id: str | None = None
 
@@ -167,16 +166,11 @@ class STTOptions:
             header["X-Api-Resource-Id"] = "volc.bigasr.sauc.duration"
         else:
             header["X-Api-Resource-Id"] = "volc.bigasr.sauc.concurrent"
-        if self.app_id is None:
-            self.app_id = os.environ.get("VOLCENGINE_STT_APP_ID", None)
-            if self.app_id is None:
-                raise ValueError("VOLCENGINE_STT_APP_ID is not set")
-        if self.access_token is None:
-            self.access_token = os.environ.get("VOLCENGINE_STT_ACCESS_TOKEN", None)
-            if self.access_token is None:
-                raise ValueError("VOLCENGINE_STT_ACCESS_TOKEN is not set")
-        header["X-Api-Access-Key"] = self.access_token
-        header["X-Api-App-Key"] = self.app_id
+        if self.api_key is None:
+            self.api_key = os.environ.get("VOLCENGINE_STT_API_KEY", None)
+            if self.api_key is None:
+                raise ValueError("VOLCENGINE_STT_API_KEY is not set")
+        header["X-Api-Key"] = self.api_key
         header["X-Api-Connect-Id"] = reqid
         header["X-Api-Request-Id"] = reqid
         return header
@@ -186,9 +180,8 @@ class STT(stt.STT):
     def __init__(
         self,
         *,
-        app_id: str | None = None,
+        api_key: str | None = None,
         base_url: str = "wss://openspeech.bytedance.com/api/v3/sauc/bigmodel",
-        access_token: str | None = None,
         resource_id: str | None = None,
         model_name: str = "bigmodel",
         enable_itn: bool = False,
@@ -208,8 +201,7 @@ class STT(stt.STT):
 
         self._opts = STTOptions(
             base_url=base_url,
-            access_token=access_token,
-            app_id=app_id,
+            api_key=api_key,
             resource_id=resource_id,
             model_name=model_name,
             enable_itn=enable_itn,

{livekit_plugins_volcenginee-1.3.0 → livekit_plugins_volcenginee-1.3.1}/livekit/plugins/volcengine/tts.py

@@ -25,8 +25,7 @@ from .log import logger
 
 
 class _TTSOptions(BaseModel):
-    app_id: str
-    access_token: str | None = None
+    api_key: str
     resource_id: str | None = None
     voice: str = "zh_female_xiaohe_uranus_bigtts"
     base_url: str = "https://openspeech.bytedance.com"
@@ -60,18 +59,12 @@ class _TTSOptions(BaseModel):
         }
 
     def get_http_header(self, reqid: str | None = None) -> dict[str, str]:
-        if self.access_token is None:
-            self.access_token = os.getenv("VOLCENGINE_TTS_ACCESS_TOKEN")
-            if self.access_token is None:
-                raise ValueError("VOLCENGINE_TTS_ACCESS_TOKEN is not set")
         resource_id = self.resource_id or os.getenv(
             "VOLCENGINE_TTS_RESOURCE_ID", "seed-tts-2.0"
         )
         headers = {
             "Content-Type": "application/json",
-            "X-Api-App-Id": self.app_id,
-            "X-Api-App-Key": self.app_id,
-            "X-Api-Access-Key": self.access_token,
+            "X-Api-Key": self.api_key,
             "X-Api-Resource-Id": resource_id,
         }
         if reqid is not None:
@@ -82,8 +75,7 @@ class _TTSOptions(BaseModel):
 class TTS(tts.TTS):
     def __init__(
         self,
-        app_id: str,
-        access_token: str | None = None,
+        api_key: str,
         resource_id: str | None = None,
         voice: str = "zh_female_xiaohe_uranus_bigtts",
         speed: float = 1.0,
@@ -95,8 +87,7 @@ class TTS(tts.TTS):
         """VolcEngine TTS
 
         Args:
-            app_id (str): the app id of the tts, you can get it from the console.
-            access_token (str | None, optional): the access token of the tts, if not provided, the value of the environment variable VOLCENGINE_TTS_ACCESS_TOKEN will be used. Defaults to None.
+            api_key (str): the API key of the tts, you can get it from the new console.
             resource_id (str | None, optional): VolcEngine TTS resource id. Use `seed-tts-2.0` for Doubao TTS 2.0 voices, `seed-tts-1.0` / `seed-tts-1.0-concurr` for Doubao TTS 1.0 voices, `seed-icl-2.0` for voice cloning 2.0, and `seed-icl-1.0` / `seed-icl-1.0-concurr` for voice cloning 1.0. Defaults to None.
             voice (str, optional): the voice id used by the tts request. Defaults to `zh_female_xiaohe_uranus_bigtts`.
             sample_rate (Literal[24000, 16000, 8000], optional): the sample rate of the tts. Defaults to 24000.
@@ -109,8 +100,7 @@ class TTS(tts.TTS):
             num_channels=1,
         )
         self._opts = _TTSOptions(
-            app_id=app_id,
-            access_token=access_token,
+            api_key=api_key,
             resource_id=resource_id,
             voice=voice,
             sample_rate=sample_rate,

livekit_plugins_volcenginee-1.3.1/livekit/plugins/volcengine/version.py

@@ -0,0 +1 @@
+__version__ = "1.3.1"

livekit_plugins_volcenginee-1.3.0/livekit/plugins/volcengine/version.py

@@ -1 +0,0 @@
-__version__ = "1.3.0"