diff --git a/src/fish_audio_sdk/apis.py b/src/fish_audio_sdk/apis.py
index fad1098..dfaa180 100644
--- a/src/fish_audio_sdk/apis.py
+++ b/src/fish_audio_sdk/apis.py
@@ -8,6 +8,7 @@
     ASRRequest,
     ASRResponse,
     ModelEntity,
+    Backends,
     PackageEntity,
     PaginatedResponse,
     TTSRequest,
@@ -16,11 +17,11 @@
 
 class Session(RemoteCall):
     @convert_stream
-    def tts(self, request: TTSRequest) -> GStream:
+    def tts(self, request: TTSRequest, backend: Backends = "speech-1.5") -> GStream:
         yield Request(
             method="POST",
             url="/v1/tts",
-            headers={"Content-Type": "application/msgpack"},
+            headers={"Content-Type": "application/msgpack", "model": backend},
             content=ormsgpack.packb(request.model_dump()),
         )
 
diff --git a/src/fish_audio_sdk/schemas.py b/src/fish_audio_sdk/schemas.py
index 232223d..0bdbb75 100644
--- a/src/fish_audio_sdk/schemas.py
+++ b/src/fish_audio_sdk/schemas.py
@@ -4,6 +4,9 @@
 
 from pydantic import BaseModel, Field
 
+
+Backends = Literal["speech-1.5", "agent-x0"]
+
 Item = TypeVar("Item")
 
 
diff --git a/src/fish_audio_sdk/websocket.py b/src/fish_audio_sdk/websocket.py
index 2f444b7..41efe33 100644
--- a/src/fish_audio_sdk/websocket.py
+++ b/src/fish_audio_sdk/websocket.py
@@ -8,7 +8,7 @@
 
 from .exceptions import WebSocketErr
 
-from .schemas import CloseEvent, StartEvent, TTSRequest, TextEvent
+from .schemas import Backends, CloseEvent, StartEvent, TTSRequest, TextEvent
 
 
 class WebSocketSession:
@@ -37,9 +37,16 @@ def close(self):
         self._client.close()
 
     def tts(
-        self, request: TTSRequest, text_stream: Iterable[str]
+        self,
+        request: TTSRequest,
+        text_stream: Iterable[str],
+        backend: Backends = "speech-1.5",
     ) -> Generator[bytes, None, None]:
-        with connect_ws("/v1/tts/live", client=self._client) as ws:
+        with connect_ws(
+            "/v1/tts/live",
+            client=self._client,
+            headers={"model": backend},
+        ) as ws:
 
             def sender():
                 ws.send_bytes(
@@ -102,9 +109,16 @@ async def close(self):
         await self._client.aclose()
 
     async def tts(
-        self, request: TTSRequest, text_stream: AsyncIterable[str]
+        self,
+        request: TTSRequest,
+        text_stream: AsyncIterable[str],
+        backend: Backends = "speech-1.5",
     ) -> AsyncGenerator[bytes, None]:
-        async with aconnect_ws("/v1/tts/live", client=self._client) as ws:
+        async with aconnect_ws(
+            "/v1/tts/live",
+            client=self._client,
+            headers={"model": backend},
+        ) as ws:
 
             async def sender():
                 await ws.send_bytes(