markitdown — File-to-Markdown Conversion

markitdown — File-to-Markdown Conversion

Use markitdown (via uvx) to convert binary documents to readable markdown before analysis or ingestion.

Standard Invocation

uvx --from 'markitdown[pdf,docx,pptx,xlsx,outlook]' markitdown <input-file> --output /tmp/<basename>.md

Then read the output file:

# or use the Read tool on /tmp/<basename>.md

Supported Formats

Extension	Format
.pdf	PDF documents
.docx	Word documents
.pptx	PowerPoint presentations
.xlsx	Excel spreadsheets
.msg, .eml	Outlook/email messages

When to Use

• User provides a binary document for analysis (PDF report, Word doc, PowerPoint)
• Need to extract text content from Office files for summarisation
• Preprocessing documents before feeding to further analysis

Pattern

1. User shares a file path (e.g., /Users/bolster/Downloads/report.pdf)
2. Run: uvx --from 'markitdown[pdf,docx,pptx,xlsx,outlook]' markitdown /Users/bolster/Downloads/report.pdf --output /tmp/report.md
3. Read /tmp/report.md with the Read tool
4. Proceed with analysis on the extracted text

Note: Always write to /tmp/ to avoid cluttering the user's directories.