Open a .docx, edit the text and structure in the browser, and export back to .docx — without Word installed or a Microsoft subscription. It is built for text-level editing: fixing content, restructuring headings, updating lists and links. It is not trying to be Word, and knowing exactly where its boundaries sit will save you a bad surprise.
What the import keeps — and the underline surprise
Import runs through mammoth, which converts .docx to clean semantic HTML: headings, bold and italic, bulleted and numbered lists, hyperlinks, and images come through. One deliberate mammoth default applies here: underlined text loses its underline on import, because underlines in converted documents are too easily mistaken for links. You can re-apply underline in the editor and it will export fine — it just does not arrive from the source file. Comments and headers/footers do not import; footnotes and endnotes do, arriving as plain linked notes at the end of the document rather than anchored to the bottom of each page. The old binary .doc format is rejected — only .docx works. If you have a .doc, re-save it as .docx in any modern word processor first.
What a .docx export can and cannot encode
Export covers what the toolbar can produce: headings (H1 through H3 from the toolbar; an H4 carried in from an imported file survives too), paragraphs, bold, italic, underline, single-level bulleted and numbered lists, hyperlinks, blockquote indents, and line breaks. Three things degrade on the way out. Nested lists flatten to a single level. Tables — which do render in the editor after import — collapse into plain text in one paragraph in the exported .docx. And images are dropped from .docx export entirely, though they are preserved if you export as HTML instead. The exported file uses Word's default fonts, page size, and margins; source formatting like custom fonts and colors does not round-trip.
The practical consequence: this editor round-trips prose documents cleanly — reports, letters, meeting notes, anything that is mostly headings, paragraphs, and lists. For a document whose value is in its tables, images, or layout, edit the text here if convenient but do the final export from a full word processor, or use the HTML export as the fidelity-preserving path.
Working in the editor
The toolbar covers bold/italic/underline (with the standard Ctrl+B/I/U shortcuts), three heading levels plus plain paragraph, both list types, link insertion, and a clear-formatting button that strips inline styles and resets the block to a paragraph. A live word and character count sits in the header — the same counting logic as a dedicated word counter, updating as you type. You can also start from a blank document rather than uploading, which makes this a quick way to produce a .docx from scratch on a machine with no office suite.
Questions people ask
Which parts of this run on a server — the .docx parsing, the export, or both?
Neither. The .docx parser (mammoth) and the .docx writer both run as JavaScript in your browser tab, and the document's contents are never transmitted anywhere — there is no upload endpoint for them to go to.
Will my table survive a round trip?
It will display in the editor, but the .docx export flattens tables to a run of plain text. If the table matters, export as HTML — tables survive that path — or make your text edits here and rebuild the export in a word processor.
The word counter gives deeper stats (sentences, reading time) on plain text, and the markdown editor is the better home for documents that will live in a repo rather than a .docx.