From 2ec7eab8e2258616830227b99aeee4143a89c201 Mon Sep 17 00:00:00 2001
From: Soulter <905617992@qq.com>
Date: Sat, 27 Jun 2026 16:58:13 +0800
Subject: [PATCH] fix(provider): handle audio parts in OpenAI Responses adapter

---
 .../sources/openai_responses_source.py        |  4 +--
 tests/test_openai_responses_source.py         | 32 +++++++++++++++++++
 2 files changed, 34 insertions(+), 2 deletions(-)

diff --git a/astrbot/core/provider/sources/openai_responses_source.py b/astrbot/core/provider/sources/openai_responses_source.py
index 462c71500..e5d99e347 100644
--- a/astrbot/core/provider/sources/openai_responses_source.py
+++ b/astrbot/core/provider/sources/openai_responses_source.py
@@ -71,8 +71,8 @@ class ProviderOpenAIResponses(ProviderOpenAIOfficial):
                 if detail:
                     image_part["detail"] = detail
                 converted.append(image_part)
-            elif part_type == "input_audio":
-                converted.append(part)
+            elif part_type in {"audio_url", "input_audio"}:
+                converted.append({"type": text_type, "text": "[Audio]"})
             elif part_type == "think":
                 continue
             else:
diff --git a/tests/test_openai_responses_source.py b/tests/test_openai_responses_source.py
index a98f26330..34d348611 100644
--- a/tests/test_openai_responses_source.py
+++ b/tests/test_openai_responses_source.py
@@ -103,6 +103,38 @@ def test_chat_payload_to_responses_payload_converts_messages_and_tool_calls():
     ]
 
 
+def test_chat_payload_to_responses_payload_replaces_audio_parts_with_placeholder():
+    payload = {
+        "model": "gpt-4.1",
+        "messages": [
+            {
+                "role": "user",
+                "content": [
+                    {"type": "text", "text": "listen"},
+                    {
+                        "type": "input_audio",
+                        "input_audio": {"data": "abc", "format": "wav"},
+                    },
+                    {"type": "audio_url", "audio_url": {"url": "data:audio/wav,abc"}},
+                ],
+            },
+        ],
+    }
+
+    converted = ProviderOpenAIResponses._chat_payload_to_responses_payload(payload)
+
+    assert converted["input"] == [
+        {
+            "role": "user",
+            "content": [
+                {"type": "input_text", "text": "listen"},
+                {"type": "input_text", "text": "[Audio]"},
+                {"type": "input_text", "text": "[Audio]"},
+            ],
+        },
+    ]
+
+
 def test_build_responses_request_shares_tool_and_extra_body_handling():
     provider = _make_provider()
     payload = {