LiteLLM Audio Input Agent

Pass an MP3 file as audio input to a gpt-audio model routed through LiteLLM.

audio_input_agent.py
"""
Litellm Audio Input Agent
=========================

Cookbook example for `litellm/audio_input_agent.py`.
"""

import requests
from agno.agent import Agent
from agno.media import Audio
from agno.models.litellm import LiteLLM

# ---------------------------------------------------------------------------
# Create Agent
# ---------------------------------------------------------------------------

# Fetch the QA audio file and convert it to a base64 encoded string
url = "https://agno-public.s3.us-east-1.amazonaws.com/demo_data/QA-01.mp3"
response = requests.get(url)
response.raise_for_status()
mp3_data = response.content

# Audio input requires specific audio-enabled models like gpt-audio
agent = Agent(
    model=LiteLLM(id="gpt-audio"),
    markdown=True,
)
agent.print_response(
    "What's the audio about?",
    audio=[Audio(content=mp3_data, format="mp3")],
    stream=True,
)

# ---------------------------------------------------------------------------
# Run Agent
# ---------------------------------------------------------------------------

if __name__ == "__main__":
    pass

Run the Example

Set up your virtual environment

uv venv --python 3.12
source .venv/bin/activate

Install dependencies

uv pip install -U agno litellm requests

Set your OpenAI credentials

Use an OpenAI API key with access to the requested model. The LiteLLM SDK calls the provider directly. An existing LITELLM_API_KEY overrides provider-specific credentials, so clear it for this example.

unset LITELLM_API_KEY
export OPENAI_API_KEY="your_provider_api_key_here"

Run the example

Save the code above as audio_input_agent.py, then run:

python audio_input_agent.py

Full source: cookbook/90_models/litellm/audio_input_agent.py