diff --git a/skillclaw/protocols/openai_responses.py b/skillclaw/protocols/openai_responses.py
index 3691333..b628a61 100644
--- a/skillclaw/protocols/openai_responses.py
+++ b/skillclaw/protocols/openai_responses.py
@@ -48,6 +48,12 @@ def content_to_openai_chat_content(content: Any) -> str | list[dict[str, Any]]:
                 image_url = image_url.get("url")
             if isinstance(image_url, str) and image_url:
                 parts.append({"type": "image_url", "image_url": {"url": image_url}})
+        elif item_type in {"input_video", "video_url"}:
+            video_url = item.get("video_url") or item.get("url")
+            if isinstance(video_url, dict):
+                video_url = video_url.get("url")
+            if isinstance(video_url, str) and video_url:
+                parts.append({"type": "video_url", "video_url": {"url": video_url}})
     if not parts:
         return ""
     if all(part.get("type") == "text" for part in parts):
diff --git a/tests/test_responses_codex.py b/tests/test_responses_codex.py
index 47a12aa..9824c12 100644
--- a/tests/test_responses_codex.py
+++ b/tests/test_responses_codex.py
@@ -120,6 +120,11 @@ def test_codex_responses_multimodal_image_input_converts_to_openai_chat_content_
         }
     ]
 
+def test_codex_responses_minimax_video_input_converts_to_openai_chat_content_parts():
+    body = {"model":"MiniMax-M3","input":[{"type":"message","role":"user","content":[{"type":"input_video","video_url":"https://example.test/video.mp4"}]}]}
+    converted = openai_responses.to_openai_body(body, "fallback-model")
+    assert converted["messages"][0]["content"] == [{"type":"video_url","video_url":{"url":"https://example.test/video.mp4"}}]
+
 
 def test_codex_responses_skill_function_call_and_output_convert_to_chat_tool_roundtrip():
     body = {