Add DocConverter for legacy .doc (Word 97-2003) files

[vibeyclaw]

Summary

Adds support for converting legacy (Word 97-2003 binary format) files to Markdown, addressing #23.

Approach

The new DocConverter uses system-level tools (antiword or catdoc) to extract text from legacy .doc files, consistent with how other converters in the project handle format-specific dependencies.

Why system tools?

The legacy .doc format (OLE2/CFB binary) is complex to parse purely in Python. antiword and catdoc are well-established, widely available tools that handle this reliably.

Changes

• converters/_doc_converter.py — new DocConverter class
• converters/__init__.py — export DocConverter
• _markitdown.py — import and register DocConverter

Usage

Install a system tool first:

# macOS
brew install antiword

# Ubuntu/Debian
apt install antiword

Then use normally:

from markitdown import MarkItDown
md = MarkItDown()
result = md.convert("document.doc")
print(result.text_content)

If neither tool is installed, a MissingDependencyException is raised with clear installation instructions.

Notes

• antiword is preferred (tried first); catdoc is the fallback
• A temp file is used since both tools require a file path rather than stdin
• No new Python dependencies required

[vibeyclaw]

@microsoft-github-policy-service agree