readme.md

• Run terraform to deploy azure TTS
• Deploy the rest of the code as an azure function
• Replace secrets with your own

"this is not secure, u hard code ur api secrets!?!?!?"

idc, this is all on a free azure trial for a non prod workload used only by me.

function_app.py

import azure.functions as func
import logging
import azure.cognitiveservices.speech as speechsdk

# Azure Speech API key and region
azure_region = "XXXX"
azure_key = "XXXX"

app = func.FunctionApp(http_auth_level=func.AuthLevel.ANONYMOUS)

@app.route(route="tts")
def text_to_speech(req: func.HttpRequest) -> func.HttpResponse:
    logging.info('Python HTTP trigger function processed a request.')

    key = req.params.get('key')
    if not key:
        raise ValueError("Please pass key in the query string")
    if key != "XXXXX":
        raise ValueError("Invalid key")

    # Get text from request body or query parameters
    text = req.params.get('text')
    if not text:
        try:
            req_body = req.get_json()
        except ValueError:
            pass
        else:
            text = req_body.get('text')

    if not text:
        raise ValueError("No text provided")

    try:
        # Configure speech synthesis
        speech_config = speechsdk.SpeechConfig(subscription=azure_key, region=azure_region)
        speech_config.speech_synthesis_voice_name = 'ja-JP-NanamiNeural'
        speech_config.set_speech_synthesis_output_format(speechsdk.SpeechSynthesisOutputFormat.Ogg24Khz16BitMonoOpus)
        
        # Use pull stream instead of file output for Azure Functions
        pull_stream = speechsdk.audio.PullAudioOutputStream()
        audio_config = speechsdk.audio.AudioOutputConfig(stream=pull_stream)
        speech_synthesizer = speechsdk.SpeechSynthesizer(speech_config=speech_config, audio_config=audio_config)
        
        # Synthesize speech
        speech_synthesis_result = speech_synthesizer.speak_text_async(text).get()
        
        # Check result
        if speech_synthesis_result.reason == speechsdk.ResultReason.SynthesizingAudioCompleted:
            logging.info(f"Speech synthesized for text: {text}")
            
            # Read audio data from the stream
            audio_data = bytes(speech_synthesis_result.audio_data)
            
            return func.HttpResponse(
                audio_data,
                mimetype="audio/ogg",
                status_code=200
            )
        elif speech_synthesis_result.reason == speechsdk.ResultReason.Canceled:
            cancellation_details = speech_synthesis_result.cancellation_details
            logging.error(f"Speech synthesis canceled: {cancellation_details.reason}")
            if cancellation_details.reason == speechsdk.CancellationReason.Error:
                logging.error(f"Error details: {cancellation_details.error_details}")
            return func.HttpResponse(
                f"Speech synthesis failed: {cancellation_details.reason}",
                status_code=500
            )
    except Exception as e:
        logging.error(f"Error during speech synthesis: {str(e)}")
        return func.HttpResponse(
            f"An error occurred: {str(e)}",
            status_code=500
        )

    return func.HttpResponse(
        "Please pass text in the query string or request body",
        status_code=400
    )
  
  # httphttps://XXXX.azurewebsites.net/api/tts?key=XXXX&text=$s

host.json

{
  "version": "2.0",
  "logging": {
    "applicationInsights": {
      "samplingSettings": {
        "isEnabled": true,
        "excludedTypes": "Request"
      }
    }
  },
  "extensionBundle": {
    "id": "Microsoft.Azure.Functions.ExtensionBundle",
    "version": "[4.*, 5.0.0)"
  }
}

main.tf

provider "azurerm" {
  features {}
  subscription_id = "XXXX"
}

resource "azurerm_resource_group" "speech_service_rg" {
  name     = "speech-service-rg"
  location = "East US"
}

resource "azurerm_cognitive_account" "speech_service" {
  name                = "speechservice"
  location            = azurerm_resource_group.speech_service_rg.location
  resource_group_name = azurerm_resource_group.speech_service_rg.name
  kind                = "SpeechServices"
  sku_name            = "S0"
}

requirements.txt

# Uncomment to enable Azure Monitor OpenTelemetry
# Ref: aka.ms/functions-azure-monitor-python
# azure-monitor-opentelemetry

azure-functions
azure-cognitiveservices-speech