Skip to main content
This example demonstrates how to use Agno agents to generate streaming audio responses using OpenAI’s GPT-4o audio preview model. 
import base64
import wave
from typing import Iterator

from agno.agent import Agent, RunOutputEvent
from agno.models.openai import OpenAIResponses

# Audio Configuration
SAMPLE_RATE = 24000  # Hz (24kHz)
CHANNELS = 1  # Mono (Change to 2 if Stereo)
SAMPLE_WIDTH = 2  # Bytes (16 bits)

# Provide the agent with the audio file and audio configuration and get result as text + audio
agent = Agent(
    model=OpenAIResponses(
        id="gpt-5.2-audio-preview",
        modalities=["text", "audio"],
        audio={
            "voice": "alloy",
            "format": "pcm16",
        },  # Only pcm16 is supported with streaming
    ),
)
output_stream: Iterator[RunOutputEvent] = agent.run(
    "Tell me a 10 second story", stream=True
)

filename = "tmp/response_stream.wav"

# Open the file once in append-binary mode
with wave.open(str(filename), "wb") as wav_file:
    wav_file.setnchannels(CHANNELS)
    wav_file.setsampwidth(SAMPLE_WIDTH)
    wav_file.setframerate(SAMPLE_RATE)

    # Iterate over generated audio
    for response in output_stream:
        response_audio = response.response_audio  # type: ignore
        if response_audio:
            if response_audio.transcript:
                print(response_audio.transcript, end="", flush=True)
            if response_audio.content:
                try:
                    pcm_bytes = base64.b64decode(response_audio.content)
                    wav_file.writeframes(pcm_bytes)
                except Exception as e:
                    print(f"Error decoding audio: {e}")
print()

​
Key Features

  • Real-time Audio Streaming: Streams audio responses in real-time using OpenAI’s audio preview model
  • PCM16 Audio Format: Uses high-quality PCM16 format for audio streaming
  • Transcript Generation: Provides simultaneous text transcription of generated audio
  • WAV File Creation: Saves streamed audio directly to a WAV file format
  • Error Handling: Includes robust error handling for audio decoding

​
Use Cases

  • Interactive voice assistants
  • Real-time storytelling applications
  • Audio content generation
  • Voice-enabled chatbots
  • Dynamic audio responses for applications

​
Technical Details

The example configures audio streaming with 24kHz sample rate, mono channel, and 16-bit sample width. The streaming approach allows for real-time audio playback while maintaining high audio quality through the PCM16 format.