1import requests
2from kern.agent import Agent
3from kern.media import Audio
4from kern.models.litellm import LiteLLM
5
6# Fetch the QA audio file and convert it to a base64 encoded string
7url = "https://kern-public.s3.us-east-1.amazonaws.com/demo_data/QA-01.mp3"
8response = requests.get(url)
9response.raise_for_status()
10mp3_data = response.content
11
12# Audio input requires specific audio-enabled models like gpt-5-mini-audio-preview
13agent = Agent(
14 model=LiteLLM(id="gpt-5-mini-audio-preview"),
15 markdown=True,
16)
17agent.print_response(
18 "What's the audio about?",
19 audio=[Audio(content=mp3_data, format="mp3")],
20 stream=True,
21)
