Streamed voice demo not worning

[kun432]

Describe the bug

Streamed voice demo not worning.

Debug information

• Python: 3.12.10
• uv: 0.8.17
• Agents SDK: 0.0.3
• OpenAI SDK: 1.107.1
• OS: macOS 14.7.6

Repro steps

git clone https://github.com/openai/openai-agents-python && cd openai-agents-python
make sync
export OPENAI_API_KEY=XXXXXXXXXX
uv run python -m examples.voice.streamed.main

Just after it starts, a parse error happens without any key strokes. Also, even if I press K, recording does not start.

I added some codes to get traceback.

examples/voice/streamed/main.py --- 1/2 --- Python
14 14
15 15 from agents.voice import StreamedAudioInput, VoicePipeline
.. 16
.. 17 import traceback
16 18
17 19 # Import MyWorkflow class - handle both module and package use cases
18 20 if TYPE_CHECKING:

examples/voice/streamed/main.py --- 2/2 --- Python
171 173         except Exception as e:
172 174             bottom_pane = self.query_one("#bottom-pane", RichLog)
173 175             bottom_pane.write(f"Error: {e}")
... 176             tb = traceback.format_exc()
... 177             bottom_pane.write(f"[dim]{tb}[/dim]")
174 178         finally:
175 179             self.audio_player.close()
176 180

run again.

Expected behavior

Streamed voice "demo" should work out of the box