What is the quota limit of gemini-2.0-flash-live-001

[user2032]

Hi

is anyone know the actual quota limit of gemini-2.0-flash-live-001

My code is like this:

import asyncio
import os.path
from google import genai
from google.genai import types
import wave
from pydub import AudioSegment
import websockets.exceptions # 导入 websockets.exceptions

# Constants
MODEL_NAME = 'gemini-2.0-flash-live-001'
AUDIO_FOLDERNAME = "AUDIO"

def wave_file(filename, pcm, channels=1, rate=24000, sample_width=2):
   with wave.open(filename, "wb") as wf:
      wf.setnchannels(channels)
      wf.setsampwidth(sample_width)
      wf.setframerate(rate)
      wf.writeframes(pcm)

client = genai.Client()
model = "gemini-2.0-flash-live-001"

config = types.LiveConnectConfig(
    response_modalities=["AUDIO"],
    output_audio_transcription=types.AudioTranscriptionConfig(),
    system_instruction=types.Content(
            parts=[types.Part(text="you are a helpful assistant.")]
        ),
    speech_config=types.SpeechConfig(
        voice_config=types.VoiceConfig(
            prebuilt_voice_config=types.PrebuiltVoiceConfig(voice_name="Kore")
        ),
        language_code='en-US',
    ),)

async def main():
    n = 0
    while True:
        try:
            n += 1
            async with client.aio.live.connect(model=model, config=config) as session:
                wave_file_name  = "auduio_"+str(n)+".wav"
                path = os.path.join(AUDIO_FOLDERNAME, wave_file_name)

                message = "Hi, Do you know Elon Musk? give me 100+ words about him"
                await session.send_client_content(
                    turns={"role": "user", "parts": [{"text": message}]}, turn_complete=True
                )
                audio_data = []
                async for response in session.receive():
                    if response.data is not None:
                        audio_data.append(response.data)

                    if response.server_content.output_transcription:
                        transcription_text = response.server_content.output_transcription.text
                        print(transcription_text,end="")

                    if response.server_content.generation_complete:
                        print("Generation Complete detected......")
                        break

                audio_data = b"".join(audio_data)
                audio_segment = AudioSegment(
                    data=audio_data,
                    sample_width=2,
                    frame_rate=24000,
                    channels=1)
                os.makedirs(AUDIO_FOLDERNAME, exist_ok=True)
                audio_segment.export(path, format="wav")
                await session.close()
                async for response in session.receive():
                    if response.data is not None:
                        print(response.data)

        except websockets.exceptions.ConnectionClosedOK as e:
            print(f"WebSocket connection closed successfully: {e}")

if __name__ == "__main__":
    asyncio.run(main())

to save time, I set a break after generation_complete, instead of turn_complete.
but this will cause exceeding quota limit:
websockets.exceptions.ConnectionClosedError: received 1011 (internal error) You exceeded your current quota, please check your plan and billing details. For more information on this error, head to: h;
I check the rate limits, there is no possiablity to exceed.

I know live API is used for chatting. Saving files is not recommended, but what i want is saving the audio and listening again and again.

Thanks for your support.

[user2032]

To make sure there is only one session at the same time, i printed the closed OK messages.

[GUNAND_MAYANGLAMBAM]

Hi @user2032 , Welcome to the forum.

I tried running your code, and its running in a loop, generating multiple audio files continuously. Although the rate limit is mentioned in the documentation, it is not guaranteed. To avoid hitting the rate limit, you can modify the code to run only once instead of in a loop, or you can break the loop once the audio is saved.
You can try adding a break statement after the session is closed. For example:

await session.close()
print("Session closed")
break

[user2032]

Hi GUNAND
Thank you very much for your comment.
Is there a way to find how the limit is violated?
I can not find it on the google cloud console. there is no possibility that I violated any of the 2x limits.

I thinik there is only one session running in my code.
For each session, there is a message printed: WebSocket connection closed successfully: sent 1000 (OK); then received 1000 (OK)