tts rest api and custom tool (answer to speech)

[saxoji]

[tts rest api server code : main.py (using fastapi library)]

import uvicorn
from fastapi import FastAPI, HTTPException
from pydantic import BaseModel
import openai
import aiofiles
import os
import uuid
from pathlib import Path
from starlette.responses import StreamingResponse
import time

app = FastAPI()

authkey of yours any

REQUIRED_AUTH_KEY = "any generated key for you"

Model for the request

class TTSRequest(BaseModel):
api_key: str # openai api key
voice: str
model: str
input_text: str
auth_key: str #any generated key for you

Directory to save the mp3 files

TTS_DIR = "tts"
if not os.path.exists(TTS_DIR):
os.makedirs(TTS_DIR)

@app.post("/generate_audio/")
async def generate_audio(request: TTSRequest):
# Check the auth key
if request.auth_key != REQUIRED_AUTH_KEY:
raise HTTPException(status_code=403, detail="Invalid authentication key")

# Update the OpenAI API key for the request
openai.api_key = request.api_key
client = openai.OpenAI(api_key=request.api_key)

# Generate audio using OpenAI's TTS API
speech_file_name = f"{uuid.uuid4()}.mp3"
speech_file_path = Path(TTS_DIR) / speech_file_name

try:
    response = client.audio.speech.create(
        model=request.model,
        voice=request.voice,
        input=request.input_text
    )
    response.stream_to_file(speech_file_path)
    
    # Ensure the file is completely written
    while not os.path.exists(speech_file_path):
        time.sleep(0.1)
    
    # Create the audio data URL
    audio_data_url = f"https://your.custom.url/tts/{speech_file_name}"

    # Return the HTML audio tag
    html_audio = f'<audio controls><source src="{audio_data_url}" type="audio/mp3"></audio>'
    return {"html_audio": html_audio}

except Exception as e:
    raise HTTPException(status_code=500, detail=str(e))

@app.get("/tts/{file_name}")
async def serve_audio(file_name: str):
file_path = os.path.join(TTS_DIR, file_name)
if not os.path.exists(file_path):
raise HTTPException(status_code=404, detail="File not found")

async def iterfile():
    async with aiofiles.open(file_path, 'rb') as f:
        while True:
            chunk = await f.read(1024)  # 파일을 1KB씩 읽기
            if not chunk:
                break
            yield chunk

return StreamingResponse(iterfile(), media_type="audio/mp3")

if name == "main":
uvicorn.run(app, host="0.0.0.0", port=8000)

---

[in flowise custom tool code]
const fetch = require('node-fetch');

const url = 'https://your.custom.url/generate_audio/';
const data = {
'api_key': 'your open api key',
'voice': 'nova',
'model': 'tts-1',
'input_text': $answer_text,
'auth_key': 'REQUIRED_AUTH_KEY in the rest api main.py'
};
const options = {
method: 'POST',
headers: {
'Content-Type': 'application/json'
},
body: JSON.stringify(data)
};

try {
const response = await fetch(url, options);
const result = await response.json();
console.log(result.html_audio);
return result.html_audio;
} catch (error) {
console.error(error);
return '';
}