diff --git a/packages/markitdown/src/markitdown/converters/_rss_converter.py b/packages/markitdown/src/markitdown/converters/_rss_converter.py
index ce768b8473..c66ef971d9 100644
--- a/packages/markitdown/src/markitdown/converters/_rss_converter.py
+++ b/packages/markitdown/src/markitdown/converters/_rss_converter.py
@@ -86,7 +86,6 @@ class RssConverter(DocumentConverter):
 
     def __init__(self):
         super().__init__()
-        self._kwargs = {}
 
     def accepts(
         self,
@@ -147,20 +146,23 @@ def convert(
         # Pop our own keyword before forwarding the rest to markdownify.
         # strict=True raises RecursionError instead of falling back to plain text.
         strict: bool = kwargs.pop("strict", False)
-        self._kwargs = kwargs
         doc = minidom.parse(file_stream)
         doc.documentURI = stream_info.url or kwargs.get("url")
         feed_type = self._feed_type(doc)
 
         if feed_type == "rss":
-            return self._parse_rss_type(doc, strict=strict)
+            return self._parse_rss_type(doc, strict=strict, markdown_options=kwargs)
         elif feed_type == "atom":
-            return self._parse_atom_type(doc, strict=strict)
+            return self._parse_atom_type(doc, strict=strict, markdown_options=kwargs)
         else:
             raise ValueError("Unknown feed type")
 
     def _parse_atom_type(
-        self, doc: Document, *, strict: bool = False
+        self,
+        doc: Document,
+        *,
+        strict: bool = False,
+        markdown_options: dict[str, Any] | None = None,
     ) -> DocumentConverterResult:
         """Parse the type of an Atom feed.
 
@@ -191,6 +193,7 @@ def _parse_atom_type(
                     is_markup=is_markup,
                     base_url=self._get_field_base_url(entry, tag_name),
                     strict=strict,
+                    markdown_options=markdown_options,
                 )
                 for value, is_markup, tag_name in (
                     (entry_summary, summary_is_markup, "summary"),
@@ -242,14 +245,25 @@ def _get_atom_content(
         return text.strip(), False
 
     def _render_atom_content(
-        self, value: str, *, is_markup: bool, base_url: str = "", strict: bool = False
+        self,
+        value: str,
+        *,
+        is_markup: bool,
+        base_url: str = "",
+        strict: bool = False,
+        markdown_options: dict[str, Any] | None = None,
     ) -> str:
         """Render one Atom summary or content value as markdown."""
         if not is_markup:
             # Plain text is returned verbatim: routing it through the HTML
             # parser drops tag-shaped text such as ``<job_id>`` entirely.
             return value
-        return self._parse_content(value, base_url=base_url, strict=strict)
+        return self._parse_content(
+            value,
+            base_url=base_url,
+            strict=strict,
+            markdown_options=markdown_options,
+        )
 
     def _get_flattened_text(
         self, element: Element, tag_name: str, *, atom_text: bool = False
@@ -282,7 +296,11 @@ def _get_flattened_text(
         return " ".join(part for part in parts if part) or None
 
     def _parse_rss_type(
-        self, doc: Document, *, strict: bool = False
+        self,
+        doc: Document,
+        *,
+        strict: bool = False,
+        markdown_options: dict[str, Any] | None = None,
     ) -> DocumentConverterResult:
         """Parse the type of an RSS feed.
 
@@ -318,6 +336,7 @@ def _parse_rss_type(
                     value,
                     base_url=self._get_field_base_url(item, tag_name),
                     strict=strict,
+                    markdown_options=markdown_options,
                 )
                 for value, tag_name in (
                     (description, "description"),
@@ -336,14 +355,19 @@ def _parse_rss_type(
         )
 
     def _parse_content(
-        self, content: str, *, base_url: str = "", strict: bool = False
+        self,
+        content: str,
+        *,
+        base_url: str = "",
+        strict: bool = False,
+        markdown_options: dict[str, Any] | None = None,
     ) -> str:
         """Parse the content of an RSS feed item"""
         try:
             # using bs4 because many RSS feeds have HTML-styled content
             soup = BeautifulSoup(content, "html.parser")
             self._resolve_content_links(soup, base_url)
-            return _CustomMarkdownify(**self._kwargs).convert_soup(soup)
+            return _CustomMarkdownify(**(markdown_options or {})).convert_soup(soup)
         except RecursionError:
             if strict:
                 raise
diff --git a/packages/markitdown/tests/test_rss.py b/packages/markitdown/tests/test_rss.py
index 89d2cbb796..5cd2d949ba 100644
--- a/packages/markitdown/tests/test_rss.py
+++ b/packages/markitdown/tests/test_rss.py
@@ -2,6 +2,8 @@
 
 import io
 import sys
+import threading
+from concurrent.futures import ThreadPoolExecutor
 
 import pytest
 
@@ -81,6 +83,60 @@ def test_atom_without_namespace_is_still_supported() -> None:
     assert "Entry content." in result.markdown
 
 
+class _SynchronizedRssConverter(RssConverter):
+    """Pause both conversions after their call-specific options are received."""
+
+    def __init__(self) -> None:
+        super().__init__()
+        self._barrier = threading.Barrier(2)
+
+    def _parse_content(self, content: str, **kwargs) -> str:
+        self._barrier.wait(timeout=5)
+        return super()._parse_content(content, **kwargs)
+
+
+def test_concurrent_conversions_keep_call_specific_markdown_options() -> None:
+    embedded = "data:image/gif;base64,R0lGODlhAQABAIAAAAAAAP///w=="
+    feed = _feed_with_body(
+        "atom-content",
+        f'<h2>Section</h2><ul><li>item</li></ul><img src="{embedded}" alt="Pixel"/>',
+    )
+    converter = _SynchronizedRssConverter()
+    stream_info = StreamInfo(extension=".atom")
+
+    def convert(keep_data_uris: bool, heading_style: str, bullets: str) -> str:
+        return converter.convert(
+            io.BytesIO(feed),
+            stream_info,
+            keep_data_uris=keep_data_uris,
+            heading_style=heading_style,
+            bullets=bullets,
+        ).markdown
+
+    with ThreadPoolExecutor(max_workers=2) as pool:
+        a = pool.submit(convert, True, "atx", "+")
+        b = pool.submit(convert, False, "underlined", "*")
+
+    a_md = a.result()
+    b_md = b.result()
+
+    # Dimensions are asserted together so a leak in any one of them fails the test,
+    # not just keep_data_uris (see the PR review note about single-dimension coverage).
+
+    # keep_data_uris
+    assert f"![Pixel]({embedded})" in a_md
+    assert "![Pixel](data:image/gif;base64...)" in b_md
+
+    # heading_style: atx ("##") on A, setext/underlined on B
+    assert "## Section" in a_md
+    assert "## Section" not in b_md
+    assert "Section\n------" in b_md
+
+    # bullets: "+" on A, "*" on B
+    assert "+ item" in a_md
+    assert "* item" in b_md and "+ item" not in b_md
+
+
 def test_atom_ignores_elements_from_other_namespaces() -> None:
     feed = b"""<a:feed xmlns:a="http://www.w3.org/2005/Atom" xmlns="urn:example:other">
   <title>Unrelated title</title>