How to use
- Drop an old Word file (
.doc, the Word 97-2003 format) on the area above, or click to choose one. - Check the number of paragraphs, tables and page breaks that were read, and the text preview.
- Press Save as DOCX. The DOCX is built, re-opened to check that the text is in order, and then downloaded. If you only need the text, use Save as TXT.
What is carried over and what is dropped
| Carried over | Dropped |
|---|---|
| Body text (Korean, English, symbols) | Character formatting: bold, italic, size, color, font |
| Paragraph breaks, line breaks, tabs | Alignment, indents, list numbering styles |
| Tables (rows, cells and their text) | Merged cells, column widths, border styles, tables inside tables |
| Page breaks | Pictures, shapes, charts, equations, text boxes |
| Field results (hyperlink text, page numbers…) | Footnotes, endnotes, headers, footers, comments, tracked changes |
In the DOCX, all text is plain paragraphs in one font (Malgun Gothic, 11 pt). If you need formatting, open the DOC in Word, Hancom Office or LibreOffice and use Save As to make a DOCX.
How the text is recovered
A DOC is an OLE compound file holding a WordDocument stream (the text and header information) and a 0Table or 1Table stream (table of structures).
- The header (FIB) tells whether the file is encrypted, which Table stream is used, the number of body characters (
ccpText) and where the piece table (CLX) is. - The piece table splits the text into pieces and says where and how each piece is stored. A piece is either 1-byte characters (CP1252, mostly English) or UTF-16 (Korean and others). Fast-saved documents keep pieces in a scrambled order, so the pieces must be joined as the table says.
- The text is cut at the body length, which removes the footnote and header areas.
- Reserved characters in the text give the structure.
| Character | Meaning |
|---|---|
Carriage return (\r) |
End of paragraph |
\x07 |
End of a table cell (the end of a row uses the same character) |
\x0c |
Page break |
\x0b |
Line break inside a paragraph |
\x13 … \x14 … \x15 |
Field: instruction … result … end. Only the result is kept |
\x01, \x08 |
Picture or shape slot, skipped and only counted |
For example, a row with cells A and B and a row with C and D appear in the text as A\x07B\x07\x07C\x07D\x07\x07. The second consecutive \x07 ends the row, giving a 2 × 2 table (A B / C D).
Where table rows end
When a cell is empty, “an empty cell” and “the end of a row” cannot be told apart from the text alone. The tool first reads the paragraph properties of the document (the table row-end marks) to split rows exactly. For documents where that information cannot be read, it assumes a regular table with the same number of cells in each row and guesses from the text pattern; the page tells you when this happens. Tables with irregular rows can come out shifted, so check the result.
Files that cannot be opened
- Password-protected documents and the old Word 6/95 formats
- Files named
.docthat are really DOCX or RTF (the tool tells you which) - Damaged files
Related tools
To read a modern Word file use DOCX Viewer; to pull out only the text, DOCX to Text; for the other direction, PDF to Word.
The file is processed in your browser and never sent to a server.