Audio Input Agent

Send an MP3 to gpt-audio through a local LiteLLM proxy with LiteLLMOpenAI.

Download an MP3 and pass it as audio input to a gpt-audio model served by a local LiteLLM proxy.

audio_input_agent.py

"""
Please first install litellm[proxy] by running: uv pip install 'litellm[proxy]'

Before running this script, you need to start the LiteLLM server:

litellm --model gpt-audio --host 127.0.0.1 --port 4000
"""

import requests
from agno.agent import Agent, RunOutput  # noqa
from agno.media import Audio
from agno.models.litellm import LiteLLMOpenAI

# ---------------------------------------------------------------------------
# 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

# Provide the agent with the audio file and get result as text
# Note: Audio input requires specific audio-enabled models like gpt-audio
agent = Agent(
    model=LiteLLMOpenAI(id="gpt-audio"),
    markdown=True,
)
agent.print_response(
    "What is in this audio?", audio=[Audio(content=mp3_data, format="mp3")], stream=True
)

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

if __name__ == "__main__":
    pass

Run the Example

1. Set up your virtual environment

   Mac

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

   uv venv --python 3.12
   .venv\Scripts\activate
   
2. Install dependencies

   uv pip install -U agno "litellm[proxy]" openai requests
   
3. Export your API keys

   Mac/Linux

   export LITELLM_API_KEY="your_litellm_api_key_here"
   export OPENAI_API_KEY="your_openai_api_key_here"
   

   $Env:LITELLM_API_KEY="your_litellm_api_key_here"
   $Env:OPENAI_API_KEY="your_openai_api_key_here"
   
4. Start LiteLLM

   Start the audio-capable local proxy on port 4000:

   litellm --model gpt-audio --host 127.0.0.1 --port 4000
   
5. 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_openai/audio_input_agent.py