Convert documents to Markdown
Upload a document and download its Markdown version, with the headings, lists and tables where they belong.
Which formats go in
| Format | Extensions |
|---|---|
| Word | .docx, .docm |
| PowerPoint | .pptx, .pptm, .ppsx, .ppsm |
| Excel | .xlsx, .xlsm, .xls |
| OpenDocument | .odt, .ods, .odp |
| Rich Text Format | .rtf |
| EPUB | .epub |
| CSV | .csv |
.pdf |
The old binary formats — Word 97's .doc, PowerPoint's .ppt and Excel's
.xlsb — are missing on purpose: there is no free library that reads them
reliably, and we would rather not offer a format only for your upload to fail
afterwards. The fix takes a minute: open the document in Word, PowerPoint or
Excel and save it as .docx, .pptx or .xlsx.
What survives the conversion
This is not a plain text dump. The structure of the document comes across too:
- Headings, with their level. In Word and OpenDocument they are read from the STYLE applied to the paragraph, not from the font size. That distinction matters: a document laid out with big bold text has no real headings, and it will have none here either. If you want headings, apply the "Heading 1", "Heading 2"… styles.
- Lists, bulleted and numbered, nested levels included.
- Tables, as Markdown tables. A cell holding several paragraphs is folded into one line, because a Markdown table cannot hold more.
- Bold, italic and strikethrough, and links with their target.
- Images, which come out in a folder of their own, linked from the text.
- Word footnotes, at the end of the document with a numbered marker.
How the result arrives
If the document had no images, the download is a single .md file. If it did,
you get a ZIP with the .md and an imagenes folder beside it. Keep them
together: the Markdown links every image by that relative path, so splitting the
files up leaves the images broken.
Before downloading anything you can press "Preview" and read the Markdown exactly as it is, unformatted — which is exactly what the file will contain.
Two cases worth understanding
A PDF is a reconstruction. A PDF does not store paragraphs: it stores letters with coordinates. Here words are grouped into lines, lines into paragraphs by the vertical gap, and a short line in noticeably larger type is taken to be a heading. It gets it right most of the time and struggles with multi-column or heavily designed layouts. That is why the result always carries a warning. And it does not run OCR: a scanned PDF is an image, and an image has no text to pull out.
A spreadsheet becomes a table. An Excel file or a CSV comes out as one Markdown table per sheet, with the sheet name as a heading when there is more than one. Numbers and dates are written as you see them on screen, following your regional settings.
Options
- The first row is the headings: spreadsheets and CSV only. Ticked, the first row heads the table; cleared, columns are named A, B, C… and every row is data.
- Extract the images: clear it when you only want the text. The download
becomes a single
.mdand you are told the document had images. - Include the speaker notes: presentations only. Notes come out as a quote under each slide, so they are never mistaken for what was on screen.
- How to separate PDF pages: a "Page N" heading helps you find your way around a long PDF; a horizontal rule is less intrusive; with no separator at all the text flows from one page into the next like a written document.
Frequently asked questions
Why does my document come out with no headings at all?
Because its headings are not headings: they are ordinary paragraphs in big bold type. Word tells the two apart and so does this tool. Apply the "Heading 1", "Heading 2"… styles to the original and convert it again.
Is my document stored on the server?
No. The file is processed in memory and dropped when the request ends: nothing is written to disk and no copy is kept. That rule is the same for all 34 tools on the site.
Can I convert an old `.doc` file?
Not directly. Open it in Word and save it as .docx; uploading it to any modern
word processor and exporting works too. It is the only honest way to do it well:
there is no free reader for the 1997 binary format.
What happens to Excel formulas?
The RESULT of the formula is converted — what the cell displays — not the formula itself. Markdown has no way to represent a calculation, so what is kept is the value.
What if the document is password protected?
If it is an open password, the file is genuinely encrypted and nothing can be done without it: you get a clear message saying so. Remove it from the program that created the file and upload it again.
Is this useful for feeding documents to an AI model?
Yes, it is one of the most common uses: Markdown keeps the structure of the document in plain text, which is exactly what text-based tooling needs — from a language model to a static site generator — at a fraction of the original size.