https://livekit.io logo
Join Slack
Powered by
This message was deleted.
# ask-ai
s
This message was deleted.
t
Thanks for reaching out, Rahul. I'm an AI assistant and I'm here to help. Give me a moment while I research this topic.
It looks like the error is happening because the 
input_audio_transcription
parameter needs to be a properly structured instance of the 
InputAudioTranscription
class when using the specified API version. This class requires at least the 
model
field to be set to a valid transcription model. To help you fix this, please make sure that 
input_audio_transcription
is created as an instance of 
InputAudioTranscription
with the 
model
field set to 
"whisper-1"
, which is the supported transcription model. Here's a simple example to guide you:
Copy code
from livekit.plugins.openai import realtime
from openai.types.beta.realtime.session import InputAudioTranscription

# Create the InputAudioTranscription instance with the required model
input_audio_transcription = InputAudioTranscription(model="whisper-1")

session = AgentSession(
    llm=realtime.RealtimeModel.with_azure(
        azure_deployment="<model-deployment>",
        api_version="2024-10-01-preview",
        input_audio_transcription=input_audio_transcription,
    ),
)
If you'd rather disable input audio transcription completely, you can simply set 
input_audio_transcription
to 
None
in the 
RealtimeModel
configuration. Please don’t hesitate to reach out if you have any questions or need a hand with this—I’m here to help! Sources: livekit-agents/livekit/agents/llm/realtime.py | Azure OpenAI Realtime API and LiveKit | LiveKit Docs | examples/src/realtime_turn_detector.ts
b
For input audio transcription in azure which is the model that has to be used is it gpt-4o-mini-transcribe or whisper-1
t
For input audio transcription in Azure OpenAI through LiveKit, the model to use is "whisper-1". This is the supported transcription model for Azure-specific configurations. If you have any questions or need a bit more help, please don’t hesitate to reach out—I’m here for you! Sources: plugins/openai/src/realtime/api_proto.ts | livekit-plugins/livekit-plugins-openai/livekit/plugins/openai/stt.py
b
Why does this error only throw up during livekit cloud and during dev
t
I'd love to help you get to the bottom of this! Could you please share a bit more about your setup? Specifically: • Are there any differences in the configuration or environment variables between your development and LiveKit Cloud environments? • How are you configuring 
input_audio_transcription
in your code? • Which plugin or service are you using (for example, OpenAI, Google, etc.)? Having these details will really help us understand why the error is happening only in LiveKit Cloud. Thanks so much for sharing! Sources: livekit-plugins/livekit-plugins-aws/livekit/plugins/aws/experimental/realti... | agent/main.py
b
Locally no errors but when I connect to livekit cloud this error I am facing
t
It looks like the error is happening in LiveKit Cloud but not on your local setup because the cloud environment has stricter validation and configuration requirements. Don’t worry, I’m here to help you work through some common reasons and how to fix them: 1. Environment Variables: Please double-check that all the necessary variables (like 
OPENAI_API_KEY
, 
GOOGLE_CLOUD_PROJECT
, 
GOOGLE_CLOUD_LOCATION
) are properly set in LiveKit Cloud. Sometimes, these might be automatically resolved on your local machine, but the cloud needs them explicitly configured. 2. Input Audio Transcription Configuration: Make sure that 
input_audio_transcription
is either set to 
None
if you want to disable it, or correctly set up with a valid model. For Azure OpenAI, the model should be 
"whisper-1"
. 3. Authentication: Check that the API key you’re using in the cloud has the right permissions for transcription services. 4. Backend-Specific Settings: If you’re using Azure OpenAI, confirm the deployment includes the correct 
api_version
and 
azure_endpoint
. These little differences between your local and cloud setups often cause hiccups like this. If you want, I’m happy to help you go through these settings step-by-step to get everything running smoothly. Sources: livekit-plugins/livekit-plugins-baseten/livekit/plugins/baseten/stt.py
Open in Slack
PreviousNext
Powered by