Hey there! To get GPT-4o-mini up and running on Azure as your LLM, just follow these simple steps: 1. First, install the LiveKit OpenAI Plugin:

pip install "livekit-agents[openai]~=1.0"

2. Next, set your environment variables: ◦ `AZURE_OPENAI_API_KEY`: Pop in your Azure OpenAI API key here. ◦ `AZURE_OPENAI_ENDPOINT`: This should look like

https://{your-resource-name}.<http://openai.azure.com|openai.azure.com>

. ◦ `OPENAI_API_VERSION`: Use the API version for Azure OpenAI, like

2024-05-01-preview

. 1. Then, initialize your LLM with the Azure settings:

from livekit.plugins.openai import LLM

azure_llm = LLM.with_azure(
    model="gpt-4o-mini",
    azure_deployment="your-gpt-4o-mini-deployment-name",  # This should match your Azure deployment name
    azure_endpoint=os.getenv("AZURE_OPENAI_ENDPOINT"),    # Optional if you’ve set the env variable
    api_version=os.getenv("OPENAI_API_VERSION"),          # Optional if you’ve set the env variable
    api_key=os.getenv("AZURE_OPENAI_API_KEY"),            # Optional if you’ve set the env variable
    temperature=0.8,
)

2. Finally, plug it into your agent like this:

session = AgentSession(
    llm=azure_llm,
    # ... and don’t forget your other parts like stt, tts, vad, etc.
)

Just double-check that the deployment name (

azure_deployment

) exactly matches what you named your

gpt-4o-mini

model in the Azure portal. If you hit any snags or want a hand with anything else, just give me a shout—I’m here to help! Sources: OpenAI LLM integration guide | LiveKit Docs | OpenAI compatible LLMs | LiveKit Docs | Azure integration guide | LiveKit Docs | Large language model (LLM) integrations | LiveKit Docs