Convert Word, PPT, PDF to Markdown in 4.7 ms – 14 Formats Supported

Why "File‑to‑Markdown" Is a Real Problem

People often need to copy content from Word, PPT, or Excel into Markdown notes, reports, or resumes, spending half a day just cleaning up formatting, especially when dealing with long, messy documents.

Traditional solutions are either manual copy‑paste, which is time‑consuming and error‑prone, or online converters that require uploading files—problematic for large or sensitive documents.

Anydoc addresses the "format fragmentation" pain point by accepting any of the common office formats (.docx, .pptx, .xlsx, etc.) and outputting a single, clean Markdown representation.

How Anydoc Achieves Speed and Stability

Compared with similar tools (Pandoc, LibreOffice, markitdown), Anydoc differentiates itself in three ways:

Full format coverage : Official benchmarks show support for 14 formats, the only tool covering all of them; Pandoc supports 5, markitdown 6.

Extreme speed : Median conversion time is about 4.7 ms, whereas LibreOffice takes roughly 1,129 ms—two orders of magnitude slower.

Stable output : All formats are parsed into a unified document model, then rendered to Markdown, ensuring that fixes to one format (e.g., docx tables) automatically benefit others (e.g., rtf, odt, xlsx).

Local PDF Parsing Without Cloud Services

Many converters fail on scanned PDFs, requiring OCR services or external uploads. Anydoc parses text‑based PDFs locally using its built‑in pdf‑inspector component, avoiding third‑party servers.

This is valuable for sensitive content such as contracts or quotations, which can be processed entirely on the user's machine.

Anydoc also detects file format by inspecting internal signatures (PDF header, RTF groups, OLE streams, ZIP types) rather than relying on file extensions, allowing correct parsing even when extensions are wrong.

Usage: One‑Liner CLI or Library Integration

Anydoc offers multiple entry points, from zero‑setup CLI to deep integration in code.

CLI example (single command):

npx @firecrawl/anydoc report.docx npx @firecrawl/anydoc slides.pptx -o slides.md

Node.js and Python bindings are provided. Python example:

import anydoc
markdown = anydoc.to_markdown("report.docx")

For developers building agents or AI tools, Anydoc includes an Agent Skill that can be added with: npx skills add firecrawl/anydoc After adding, tools like Claude Code, Codex, or Cursor can directly read office documents.

Things to Keep in Mind

Anydoc converts textual content only; scanned‑image PDFs require the hosted Firecrawl API for OCR.

The tool focuses on extracting clean Markdown, not recreating pixel‑perfect layouts. Complex structures such as merged cells, footnotes, or comments are retained in Markdown but not rendered with original visual fidelity.

The project launched in early August 2026, has rapidly growing stars, but the ecosystem is still early; edge‑case formats may need user feedback via GitHub Issues.

Who Should Use It

Writers of public accounts or technical documentation who want to turn client PPTs/Word drafts into clean Markdown quickly.

AI tool developers needing agents to read various office documents via the provided Agent Skill.

Data/knowledge managers who wish to batch‑convert legacy Office files into searchable Markdown archives.

Anyone handling sensitive content who prefers local conversion without uploading to third‑party servers.

The deciding question is whether you frequently need to turn diverse office documents into a unified Markdown format; if so, Anydoc can save a substantial amount of manual effort.

github.com/firecrawl/anydoc

GitHub repository:

github.com/firecrawl/anydoc