From dc2cf01868f8c632167449b895a062f8e28a86df Mon Sep 17 00:00:00 2001
From: Andy <152056884+jesko2004@users.noreply.github.com>
Date: Sat, 19 Sep 2026 23:18:35 +0800
Subject: [PATCH] Fix charset parsing of quoted HTTP Content-Type parameters

---
 .../markitdown/src/markitdown/_markitdown.py  | 15 ++---
 .../tests/test_response_content_type.py       | 55 +++++++++++++++++++
 2 files changed, 63 insertions(+), 7 deletions(-)
 create mode 100644 packages/markitdown/tests/test_response_content_type.py

diff --git a/packages/markitdown/src/markitdown/_markitdown.py b/packages/markitdown/src/markitdown/_markitdown.py
index 888d1e6260..3e533a02d3 100644
--- a/packages/markitdown/src/markitdown/_markitdown.py
+++ b/packages/markitdown/src/markitdown/_markitdown.py
@@ -544,13 +544,14 @@ def convert_response(
         charset: Optional[str] = None
 
         if "content-type" in response.headers:
-            parts = response.headers["content-type"].split(";")
-            mimetype = parts.pop(0).strip()
-            for part in parts:
-                if part.strip().startswith("charset="):
-                    _charset = part.split("=")[1].strip()
-                    if len(_charset) > 0:
-                        charset = _charset
+            content_type = response.headers["content-type"]
+            mimetype = content_type.split(";", 1)[0].strip()
+            # A semicolon inside a quoted parameter is not a delimiter.
+            message = Message()
+            message["content-type"] = content_type
+            _charset = message.get_param("charset")
+            if isinstance(_charset, str) and _charset.strip():
+                charset = _charset.strip()
 
         # If there is a content-disposition header, get the filename and possibly the extension
         filename: Optional[str] = None
diff --git a/packages/markitdown/tests/test_response_content_type.py b/packages/markitdown/tests/test_response_content_type.py
new file mode 100644
index 0000000000..2281701cc9
--- /dev/null
+++ b/packages/markitdown/tests/test_response_content_type.py
@@ -0,0 +1,55 @@
+import io
+
+import pytest
+import requests
+
+from markitdown import MarkItDown, StreamInfo
+
+
+def _response(content_type: str | None, body: bytes) -> requests.Response:
+    response = requests.Response()
+    response.status_code = 200
+    response.url = "https://example.com/document.txt"
+    response.raw = io.BytesIO(body)
+    if content_type is not None:
+        response.headers["Content-Type"] = content_type
+    return response
+
+
+@pytest.fixture(scope="module")
+def converter() -> MarkItDown:
+    return MarkItDown(enable_plugins=False)
+
+
+@pytest.mark.parametrize(
+    "content_type",
+    [
+        'text/plain; charset=iso-8859-1; profile="urn:example;charset=utf-8"',
+        'text/plain; charset=iso-8859-1; profile="urn:example;charset=ascii;v=1"',
+        'text/plain; profile="urn:example;charset=utf-8"; charset=iso-8859-1',
+        'text/plain; charset="iso-8859-1"',
+        "text/plain; charset=iso-8859-1",
+    ],
+)
+def test_quoted_parameter_cannot_override_charset(converter, content_type):
+    response = _response(content_type, b"Caf\xe9")
+
+    assert converter.convert_response(response).markdown == "Caf\u00e9"
+
+
+@pytest.mark.parametrize("content_type", [None, "text/plain", 'text/plain; charset=""'])
+def test_response_without_charset_still_converts(converter, content_type):
+    assert (
+        converter.convert_response(_response(content_type, b"hello")).markdown
+        == "hello"
+    )
+
+
+def test_explicit_stream_info_charset_overrides_header(converter):
+    response = _response("text/plain; charset=utf-8", b"Caf\xe9")
+
+    result = converter.convert_response(
+        response, stream_info=StreamInfo(charset="iso-8859-1")
+    )
+
+    assert result.markdown == "Caf\u00e9"