Convert HWP to HTML or plain text in your browser

How to use it

1. Select your source documents and choose HTML for a formatted reading copy or plain text for extraction.
2. Run the selected conversion, then check the result summary for skipped elements and changes to table representation.
3. Download and inspect the HTML in a browser or the TXT in a text editor; use the browser’s print preview separately if a printed copy is needed.

About the formats

An HTML reading copy contains document markup, inline CSS and supported embedded images. It can be handled as one file rather than a page plus a separate image directory. The file includes print styling, but paper size, scaling, margins and the print engine still influence the resulting pages. TXT is a different product: UTF-8 text without font names, colours or visual table boundaries. Paragraph breaks and tab separators remain useful for searching and further processing. Choose based on the next operation rather than on file size alone. An image-heavy HTML file can be larger than expected because the picture bytes are encoded inside it; a small TXT file may omit exactly the diagrams that explain the text.

Frequently asked questions

Does this tool export a PDF file directly?

PDF is not a download format here. You can open the generated HTML and use a browser’s print dialog, selecting a PDF destination if your device provides one. Inspect that dialog’s preview, especially wide tables and page breaks, before treating the printout as a finished copy.

Are pictures saved next to the HTML?

Supported embedded pictures are written into the HTML as data URIs. There is no companion assets directory for this output. A linked source image that is unavailable to the parser cannot be recovered merely by packaging the HTML; the result summary is where such limitations become visible.

Can I import the TXT as a spreadsheet?

With the table-as-TSV option enabled, tabs can help with a simple rectangular table, but the file is not a guaranteed spreadsheet dataset. Merged cells appear once and covered positions are omitted, so some rows have fewer fields. Check headings, row lengths, decimal values and totals before assigning columns or running calculations.

Why is a page header missing?

These outputs focus on the document body and notes, so page headers and footers are excluded. A document identifier placed only in a header will therefore need separate attention. Keep the original attachment when that identifier, a printed page number or a repeated notice matters to your task.

Can the HTML run scripts from the input document?

The emitter writes a new HTML document with escaped text and controlled markup rather than copying arbitrary source HTML into it. Its output contains no script tags or external resource loads. Links are still links: choosing to follow one is a separate browser navigation under your control.

Why does changing the extension not fix an error?

The parser checks file contents, not just the suffix. Renaming an unrelated file to HWPX does not make it a valid package, and protected documents remain protected. If the attachment is incomplete, download it again from its source; if it is restricted, ask the sender for an appropriate export.

Inspect before printing

First view the HTML at a comfortable width and compare a long table with the source. Then open print preview and check the first page, a page containing a wide table, and the final notes. A browser may split content differently after you change orientation or scaling. A constructed example is a form with a signature area below a table: if the signature moves to another sheet, revise the print settings or choose another document workflow. The presence of an @page rule does not certify the layout of a submitted form.

Read tab-separated rows carefully

A tab separates text fields; it does not record the original visual width of a cell. In a constructed expense example, a category name spans two rows while each row holds a different amount. TXT preserves the category text once, so copying the lines into a spreadsheet without review can shift the apparent relationship between labels and numbers. Begin with a few rows, count the resulting fields, and only then decide how to represent the group. Preserve a separate reading copy so that later corrections can be checked against the surrounding explanation.

Check what text extraction can see

Words drawn inside a scanned page are image pixels, not paragraph text. The TXT emitter does not recognise those pixels as characters. Equations can remain as source script text, and unsupported objects can become placeholders or warnings. Search the result for a distinctive heading and for material near the end, but also look at the source for charts or screenshots whose meaning is absent from the extracted words. An empty text result is possible for an image-only document even when the HTML copy contains a visible picture.

Keep a reading copy useful over time

Give the downloaded copy a meaningful filename and retain the original separately if you need to trace a quotation later. HTML is convenient to move because supported images travel inside it, but that convenience is not a content-verification service. Check dates, units, citations and the reading order of a multi-column source. The current corpus check compares HTML body text with TXT from the same document model and checks well-formed output on 41 supported samples. It does not measure every browser’s printing behaviour or establish that a converted form will be accepted by its recipient.

Other options

Use the viewer when you only need to read a document inside this page. Choose Markdown when the next destination expects lightweight markup and relative picture paths. For this tool, the HTML option is suited to a self-contained reading copy and the TXT option to deliberate text extraction, with different checks for each.