Server-sideDeleted in 30 minutes

Convert Word to TXT

This conversion is deliberately destructive. A .docx is a zip full of XML describing fonts, styles, revision history and layout; a .txt is a sequence of characters. Everything in the first that is not a character is discarded, and that is the entire point.

Use this without the search next time. Prathom Workbench puts Prathom's tools in your toolbar.

Add to Chrome — free

Drop your DOCX file here, or click to browse

Up to 50 MB. Deleted automatically after 30 minutes.

What it does

Table rows come out tab-separated, ready for a spreadsheet or a script
Code and indented blocks keep their leading whitespace
Opens in any editor, on any system, forever
No watermark and no sign-up

How to use Word to TXT

1
Add the DOCX
Anything Word or LibreOffice wrote. Files still named .doc are the older binary format and work here too, though .docx is what this page describes.
2
Convert
The document is opened, its text content is read in reading order, and that text is written out with no markup around it.
3
Check the encoding if the text is not English
The output is UTF-8. Most editors detect that correctly; a few older Windows tools assume a regional codepage and will show accented characters as pairs of symbols. Reopening with UTF-8 selected fixes it.

How it works

The document is opened and its text is written out in reading order — body paragraphs, table cells row by row, list items with their numbers, in the sequence a person reading the page would encounter them.

Nothing is interpreted or reformatted. There is no attempt to draw the table with ASCII borders, mark headings with underlines, or represent bold with asterisks. Anything like that would be a guess about what you intended to do next.

Why the result is so much smaller

A 6.8 KB document became a 209 byte text file in testing — a factor of thirty-three, on a document whose actual words are a couple of paragraphs.

That ratio is not compression. Almost none of a .docx is your writing. It is a zip archive containing a style catalog, a fonts table, document properties, revision identifiers, relationship maps and a great deal of XML scaffolding, and the words are a small fraction of the bytes. Strip everything that describes how the words look and very little is left.

Where this is the right tool

Feeding text into something that reads text. Search indexes, scripts, diff tools, language models, spreadsheet imports. All of them want characters, and handing them a zip full of XML means they either fail or return gibberish.

Comparing two versions of a document. Word's own comparison works on formatting as well as content, so a font change registers as a difference. Two text files diffed line by line show only what was actually written differently.

Archiving something that must still open in fifty years. A plain text file has no version, no dependencies and no format to become obsolete.

Where it is not

If the document is going to a person rather than a program, this is almost always the wrong conversion — you are sending them something that has lost its structure, and structure is how a reader navigates a long document. PDF keeps the appearance exactly; HTML keeps the headings, the tables and the links as real things. Plain text is for machines, and it is very good at that.

Examples

A document with a table in it

report.docx — headings, bold runs, a 3x3 table, 6.8 KB

report.txt — 209 bytes

Thirty-three times smaller, and the table survives in a genuinely usable form — "Region", "Q3" and "Q4" arrive on one line separated by tab characters, then a line per row. That is exactly what a spreadsheet paste or a `cut -f2` expects, which is why this conversion is common in scripting.

A document full of formatting

contract.docx — numbered clauses, a linked reference, italic definitions

contract.txt — clause numbers as literal text, no link

The list numbers come through, because they are written into the text as "1." and "2." characters. The link does not: the words that were the link remain and the address behind them is gone, with nothing left to indicate it was ever there. If the URLs matter, convert to HTML instead.

Frequently asked questions

What exactly is lost?

Every visual property — bold, italic, fonts, sizes, colors, alignment, page breaks, headers and footers. Also images, charts, comments, tracked changes and footnote markers. Heading levels go too: a heading arrives as an indented line with nothing to say whether it was a chapter title or a sub-sub-section. What remains is the words, in reading order.

Do tables survive?

In a more useful form than you might expect. Each row becomes one line and each cell is separated by a tab character, so a three-column table comes out as three tab-separated fields per line. That is a real data format — import it into a spreadsheet as tab-delimited, or process it with any command-line tool. What is lost is merged cells, which flatten, and any formatting inside the cells.

Why is there a strange character at the very start of my file?

That is a byte-order mark, three bytes that declare the file as UTF-8. Most software hides it. Some — spreadsheet importers, older parsers, a few programming languages reading the first line — show it as a stray symbol or quietly include it in your first field. If it causes trouble, most editors have a "UTF-8 without BOM" save option that removes it.

Can I convert the text file back into a Word document?

You can produce a Word document from it, but nothing that was discarded comes back. The result is your text in a default font with no headings, no table structure and no styling — a new document that happens to contain the same words. Keep the original .docx if the formatting has any value.