diff --git a/server/main.py b/server/main.py
index 2a2ca91..ccbae2c 100644
--- a/server/main.py
+++ b/server/main.py
@@ -10,9 +10,11 @@
 from fastapi import (
     FastAPI,
     UploadFile,
+    Body,
 )
 from pydantic import BaseModel
 from fastapi.responses import StreamingResponse
+
 from TTS.tts.configs.xtts_config import XttsConfig
 from TTS.tts.models.xtts import Xtts
 from TTS.utils.generic_utils import get_user_data_dir
@@ -112,9 +114,11 @@ class StreamingInputs(BaseModel):
         "ja",
     ]
     add_wav_header: bool = True
+    stream_chunk_size: str = "20"
+    decoder: str = "ne_hifigan"
 
 
-async def predict_streaming_generator(parsed_input: StreamingInputs):
+async def predict_streaming_generator(parsed_input: dict = Body(...)):
     speaker_embedding = (
         torch.tensor(parsed_input.speaker_embedding).unsqueeze(0).unsqueeze(-1)
     )
@@ -123,14 +127,15 @@ async def predict_streaming_generator(parsed_input: StreamingInputs):
     )
     text = parsed_input.text
     language = parsed_input.language
-    decoder = parsed_input.get("decoder", "ne_hifigan")
+    decoder = parsed_input.decoder
 
     if decoder not in ["ne_hifigan","hifigan"]:
         decoder = "ne_hifigan"
 
-    stream_chunk_size = int(parsed_input.get("stream_chunk_size", "20"))
+    stream_chunk_size = int(parsed_input.stream_chunk_size)
     add_wav_header = parsed_input.add_wav_header
 
+
     chunks = model.inference_stream(text, language, gpt_cond_latent, speaker_embedding, decoder=decoder,stream_chunk_size=stream_chunk_size)
     for i, chunk in enumerate(chunks):
         chunk = postprocess(chunk)