Skip to main content
PDFBase

Document Converters

3 tools

Three converters for document formats whose original software has quietly disappeared. Each one reads its source format outright — the markup, the package, the coordinates — rather than scraping text out of it, and each one tells you per file what the PDF could not reproduce.

What makes these formats hard to open today?

Each was tied to software that is gone or hidden. RTF outlived WordPad, which Windows removed in 2024, and survives mainly as the export format of systems nobody has modernised. XPS was Microsoft's answer to PDF; Windows still ships the writer that creates the files but stopped installing the viewer that reads them from Windows 10 version 1803. OXPS is the standardised successor that Windows 8 switched to, and the older viewer cannot read it at all. In every case the file is fine and the reader is missing, which is exactly the problem a converter solves.

How is a real conversion different from extracting the text?

A text extractor gives you the words with the document thrown away — no page size, no indents, no tables, no pictures, no positioning. These converters parse the format itself. For RTF that means the control-word grammar: font and colour tables, paragraph properties, cell definitions, picture data. For XPS and OXPS it means the fixed-page display list: vector paths, clipping, transforms, opacity, images and the exact coordinates of every glyph run. The distinction matters because a page of extracted text presented as a converted document is a claim the file does not support.

Why does every conversion come with a compatibility report?

Because none of these conversions is lossless, and the useful question is not whether something changed but what. PDFBase draws with the standard PDF fonts, so typefaces are substituted and line breaks can move. Windows metafile pictures cannot be drawn. Gradients are flattened to a single colour. Rather than bury that in a disclaimer, each file in the queue carries its own list of exactly which of these applied to it, and the same list is written into the manifest inside the ZIP so the record travels with the documents.

Which formats are deliberately missing from this list?

CHM help files, WordPerfect documents, Microsoft Works files, and Apple Pages and Keynote documents. All four were considered and none of them ships, for the same reason: their containers are undocumented or reverse-engineered, no credible browser-capable decoder exists for them, and there is no way to verify a converter against real files without publishing something that might quietly produce nonsense. A converter that cannot be shown to be correct is worse than no converter, because it looks like an answer.

What each document converter reads and reproduces

What each document converter reads and reproduces
FormatWhat it isLayout fidelityMain loss
RTFReflowable word-processor markupRe-laid out from the document's own page size, margins and indentsTypefaces substituted, so line breaks can move
XPSFixed-layout package of drawing instructionsExact — every mark keeps its coordinatesEmbedded fonts replaced; gradients flattened
OXPSThe standardised form of XPSExact — same engine, same coordinatesEmbedded fonts replaced; gradients flattened

Fidelity differs because the formats differ. RTF describes a document that flows, so its lines have to be broken again with the fonts available; XPS and OXPS describe a page that is already composed, so the geometry can be carried across intact.

Every tool in this category runs on your device

There is no upload step and no processing server. Open your browser’s network panel and run any of them — the only requests are for the page itself.