microsoft/markitdown · error · ValueError

Not supported tag: {tag.name}

Error message

Not supported tag: {tag.name}

What it means

During DOCX pre-processing, math tags are rewritten: oMathPara becomes a w:p paragraph of block equations and oMath becomes an inline equation. Any other tag name reaching this branch raises ValueError — it is an internal invariant check inside the OMML pre-processing pass, not a user-configurable error.

Source

Thrown at packages/markitdown/src/markitdown/converter_utils/docx/pre_process.py:96

    Args:
        tag (Tag): A BeautifulSoup Tag object representing the OMML element. Could be either "oMathPara" or "oMath".

    Raises:
        ValueError: If the tag is not supported.
    """
    if tag.name == "oMathPara":
        # Create a new paragraph tag
        p_tag = Tag(name="w:p")
        # Replace each 'oMath' child tag with its LaTeX equivalent as block equations
        for child_tag in tag.find_all("oMath"):
            p_tag.append(_get_omath_tag_replacement(child_tag, block=True))
        # Replace the original 'oMathPara' tag with the new paragraph tag
        tag.replace_with(p_tag)
    elif tag.name == "oMath":
        # Replace the 'oMath' tag with its LaTeX equivalent as inline equation
        tag.replace_with(_get_omath_tag_replacement(tag, block=False))
    else:
        raise ValueError(f"Not supported tag: {tag.name}")


def _pre_process_math(content: bytes) -> bytes:
    """
    Pre-processes the math content in a DOCX -> XML file by converting OMML (Office Math Markup Language) elements to LaTeX.
    This preprocessed content can be directly replaced in the DOCX file -> XMLs.

    Args:
        content (bytes): The XML content of the DOCX file as bytes.

    Returns:
        bytes: The processed content with OMML elements replaced by their LaTeX equivalents, encoded as bytes.
    """
    soup = BeautifulSoup(content.decode(), features="xml")
    for tag in soup.find_all("oMathPara"):
        _replace_equations(tag)
    for tag in soup.find_all("oMath"):
        _replace_equations(tag)

View on GitHub (pinned to fd239d5d2b)

Solutions

1. Re-save the document with Microsoft Word/LibreOffice (normalizes the XML) and retry
2. Upgrade markitdown — pre-processing coverage for additional tags may have been added
3. Strip the math elements or export the document without equations as a workaround
4. File an upstream issue attaching the offending document.xml snippet

try:
    result = md.convert(docx_path)
except ValueError as e:
    if "Not supported tag" in str(e):
        log.warning("nonstandard math XML; re-save document via Word/LibreOffice and retry")
        raise

When it happens

Trigger: A DOCX whose document.xml contains a math-related tag other than oMath/oMathPara reaching the replacement walker — typically from a corrupted file or a document produced by a non-standard generator with extra elements inside the math namespace.

Common situations: Third-party DOCX writers (reporting tools, converters) emitting unexpected tags in the m: namespace, or manually edited XML inside the zip.

Related errors

• Not support func %s
• Not support lim %s
• {converter} recognized the input as a potential {extension}
• File conversion failed after {len(attempts)} attempts:
• {converter} recognized the input as a potential {extension}

AI-assisted analysis of microsoft/markitdown@fd239d5d2b (2026-08-14). Data as JSON: /api/errors/aefaf7653437dc99. Report an issue: GitHub.