From ee751cb2039f2db6071a3cebb4a7ef27adbbb979 Mon Sep 17 00:00:00 2001
From: Shagun Bera <141054835+notV3NOM@users.noreply.github.com>
Date: Wed, 17 Sep 2025 10:23:41 +0530
Subject: [PATCH] feat(voice)!: migrate STT streaming to match GA Realtime API
 (#1759)

---
 src/agents/tracing/spans.py           |  1 +
 src/agents/voice/models/openai_stt.py | 13 +++++++++----
 tests/voice/test_openai_stt.py        |  6 +++---
 3 files changed, 13 insertions(+), 7 deletions(-)

diff --git a/src/agents/tracing/spans.py b/src/agents/tracing/spans.py
index d0a416b8..dbde6f9e 100644
--- a/src/agents/tracing/spans.py
+++ b/src/agents/tracing/spans.py
@@ -22,6 +22,7 @@ class SpanError(TypedDict):
         message: A human-readable error description
         data: Optional dictionary containing additional error context
     """
+
     message: str
     data: dict[str, Any] | None
 
diff --git a/src/agents/voice/models/openai_stt.py b/src/agents/voice/models/openai_stt.py
index 12333b02..f0255f24 100644
--- a/src/agents/voice/models/openai_stt.py
+++ b/src/agents/voice/models/openai_stt.py
@@ -163,11 +163,16 @@ class OpenAISTTTranscriptionSession(StreamedTranscriptionSession):
         await self._websocket.send(
             json.dumps(
                 {
-                    "type": "transcription_session.update",
+                    "type": "session.update",
                     "session": {
-                        "input_audio_format": "pcm16",
-                        "input_audio_transcription": {"model": self._model},
-                        "turn_detection": self._turn_detection,
+                        "type": "transcription",
+                        "audio": {
+                            "input": {
+                                "format": {"type": "audio/pcm", "rate": 24000},
+                                "transcription": {"model": self._model},
+                                "turn_detection": self._turn_detection,
+                            }
+                        },
                     },
                 }
             )
diff --git a/tests/voice/test_openai_stt.py b/tests/voice/test_openai_stt.py
index 12c58a22..8eefc995 100644
--- a/tests/voice/test_openai_stt.py
+++ b/tests/voice/test_openai_stt.py
@@ -115,10 +115,10 @@ async def test_session_connects_and_configures_successfully():
         assert headers.get("OpenAI-Beta") is None
         assert headers.get("OpenAI-Log-Session") == "1"
 
-        # Check that we sent a 'transcription_session.update' message
+        # Check that we sent a 'session.update' message
         sent_messages = [call.args[0] for call in mock_ws.send.call_args_list]
-        assert any('"type": "transcription_session.update"' in msg for msg in sent_messages), (
-            f"Expected 'transcription_session.update' in {sent_messages}"
+        assert any('"type": "session.update"' in msg for msg in sent_messages), (
+            f"Expected 'session.update' in {sent_messages}"
         )
 
         await session.close()