Markitdown Convert

Determine:

• the input path or stream
• whether the user wants a saved .md file or just converted output
• the desired output path, if specified

If the user does not specify an output path and clearly wants a file created, default to the source basename with a .md extension next to the source file.

2. Convert with the CLI

Save to a file:

markitdown /path/to/input.pdf -o /path/to/output.md

Write Markdown to stdout:

markitdown /path/to/input.pdf

Read from stdin when needed:

cat /path/to/input.pdf | markitdown > output.md

3. Verify the result

After conversion:

• confirm the command exited successfully
• inspect the beginning of the generated Markdown
• mention any obvious degradation or unsupported content

4. Report clearly

Tell the user:

• where the Markdown was written
• whether conversion succeeded cleanly
• any notable caveats, such as OCR limits or unsupported embedded content

Notes

• This environment has MarkItDown installed from the Microsoft GitHub repository, not the stale PyPI 0.0.2 release.
• The current upstream repo exposes an MCP server package (markitdown-mcp), but not a Codex/OpenCode skill.
• The upstream repo currently has a broken optional dependency for YouTube transcription (youtube-transcript-api~=1.0.0), so this local install excludes the YouTube extra. If the user asks to convert a YouTube URL, explain that the upstream dependency is currently not resolvable and do not pretend support exists.
• For other supported file types, use the installed CLI first before reaching for custom parsing.