from elevenlabs.client import ElevenLabs
import os
import sys
import wave

elevenlabs = ElevenLabs(
    api_key="sk_XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX",
)

audio = elevenlabs.text_to_speech.convert(
    text=sys.argv[7],
    voice_id=sys.argv[2],
    model_id="eleven_v3",
#    model_id="eleven_multilingual_v2",
    output_format="pcm_16000",
)

# Collect all chunks into memory
pcm_data = b"".join(chunk for chunk in audio)

# Write directly to WAV file
with wave.open(sys.argv[4], "wb") as wav_file:
    wav_file.setnchannels(1)            # Mono
    wav_file.setsampwidth(2)            # 16-bit PCM
    wav_file.setframerate(16000)        # Sample rate
    wav_file.writeframes(pcm_data)