Convert to Markdown
14 input formats supported — docx, odt, epub, html, latex, rst, org, csv, tsv, mediawiki, docbook, opml, ipynb and markdown itself. All powered by the Pandoc WASM engine, running privately in your browser.
Click to start converting
Go to Pandoc engine — Markdown output selected automatically
Supported input formats
| 格式 | 扩展名 | 说明 |
|---|---|---|
| Word (docx) | .docx | Headings, tables, footnotes, images, tracked changes |
| LibreOffice (odt) | .odt | Paragraph styles, master pages, embedded media |
| E-book (epub) | .epub | Chapter structure, book metadata, images |
| HTML | .html | Headings, lists, tables, links — genuine HTML reader |
| LaTeX | .tex | Math, environments, citations |
| reStructuredText | .rst | Directives, roles, tables |
| Org-mode | .org | Headlines, properties, drawers |
| CSV | .csv | Pandoc renders as GFM pipe tables |
| TSV | .tsv | Tab-separated, rendered as pipe tables |
| MediaWiki | .mediawiki | Templates, links, tables |
| DocBook | .docbook | Articles, sections, media objects |
| OPML outlines | .opml | Nested outlines → heading hierarchy |
| Jupyter notebook | .ipynb | Code cells, outputs, markdown cells |
How the conversion works
Under the hood this page runs the real Pandoc compiler, compiled to WebAssembly. When you drop a .docx, Pandoc's docx reader maps Word styles to document semantics — Heading 1 becomes a level-1 heading, list paragraphs become lists, and table grids become table nodes in an intermediate abstract syntax tree.
The Markdown writer then walks that tree and emits GitHub Flavored Markdown: pipe tables, fenced code blocks, footnotes and task lists. Because the mapping goes through a semantic tree instead of HTML heuristics, nested structures survive far more cleanly than with copy-paste converters.
Everything happens inside your browser tab in a background worker. The document is never uploaded — you can watch the Network panel stay empty while a 50 MB report converts.
Common issues and how to fix them
Legacy binary .doc files are not accepted — open them in Word or LibreOffice and save as .docx first. Password-protected files must be unlocked before conversion.
Images are extracted into a media/ folder next to the .md file. Use the ZIP download to get both together, or switch the image mode to base64 embedding when you need a single self-contained file for Obsidian or a CMS that dislikes external assets.
Comments and tracked changes have no Markdown equivalent and are resolved as plain text. Accept or reject revisions in Word first if you want a clean result. Equations are simplified to text — for math-heavy documents, convert to LaTeX instead.
Converting CSV or TSV? Pandoc treats the first row as a header and renders a pipe table, which is exactly what GitHub renders. Keep GFM enabled in the options panel for this.
Who uses this page
Note-takers moving Word archives into Obsidian or Logseq vaults, where the media/image1.png convention keeps references valid when files land in the vault.
Documentation teams migrating legacy specs into MkDocs, Docusaurus or GitHub wikis, and bloggers exporting drafts from Word to publish on static-site generators such as Hugo, Jekyll or Astro.
Convert to Markdown FAQ
+What formats can I convert to Markdown?
14 formats: docx, odt, epub, html, latex, rst, org, csv, tsv, mediawiki, docbook, opml, ipynb and markdown itself. All handled by the Pandoc engine — the de facto standard for document conversion.
+Why use Pandoc instead of Turndown or Mammoth?
Pandoc has genuine readers for each format — not HTML-based heuristics. It understands LaTeX math environments, EPUB chapter trees, Org-mode properties and DocBook semantics directly. The result is cleaner Markdown, especially for complex documents.
+Are my files uploaded to a server?
Never. Pandoc runs as a WebAssembly binary in your browser. The network tab stays empty during conversion — your documents never leave your device.
+Does it preserve images?
Yes. For docx, odt, epub and ipynb, images are extracted to a media/ folder and referenced in the Markdown. You can download everything as a ZIP, or switch to base64 embedding for a single self-contained file.
+Can I batch convert multiple files?
Yes. Drop several files at once — they queue and convert one by one, each with its own status. Download all results as a ZIP.
+Does it handle GFM (GitHub Flavored Markdown)?
Yes, GFM is on by default — pipe tables, strikethrough, task lists and fenced code blocks. Turn it off for strict CommonMark output.