tts backend

.gitignore

@@ -1,2 +1,4 @@
 **/.venv/
 **/__pycache__/
+**/audio
+*.mp3

speech-speech/backend/api.py

@@ -1,5 +1,6 @@
 from openai import OpenAI
 from fastapi import FastAPI, File, Response, Request
+from fastapi.responses import FileResponse
 from pydantic import BaseModel
 import whisper
 
@@ -49,3 +50,15 @@ async def get_next_response(request: Request, response: Response):
     print(messages)
     print(res_msg)
     return {"role": role, "content": res_msg}
+
+
+@app.post("/speak", response_class=FileResponse)
+def tts(text: str, response: Response):
+    response.headers["Access-Control-Allow-Origin"] = "*"
+    res = openAI_clinet.audio.speech.create(
+        model="tts-1",
+        voice="nova",
+        input=text,
+    )
+    res.stream_to_file("tts.mp3")
+    return "tts.mp3"