What Is an RTF File? Structure, Compatibility, and Conversion Limits

RTF began as Microsoft’s cross-application formatted-text interchange language

RTF means Rich Text Format. Microsoft shipped its first reader and writer with Word 3.0 for Macintosh in 1987, when moving a formatted document between word processors was far less predictable than it is today. The current published specification is RTF 1.9.1, written for Word 2007 in March 2008. RTF is neither a Word DOCX package nor a flat text document. It is a stream of ordinary-looking characters, braces, backslashes, and instructions that describe formatted document content. That design explains why an RTF can be inspected in a text editor even though it opens as formatted pages in Word, LibreOffice Writer, TextEdit, and other document programs.

Files commonly remain in RTF when a system exports basic letters, reports, or templates without committing to one office suite, and when an older record still needs a broadly readable editable copy. The useful promise is modest: ordinary paragraphs, font choices, bold, italic, lists, tables, pictures, and some document properties can be represented. It is not a reliable master format for the full feature set of a current word processor. A modern document with live comments, revision history, protected content controls, sophisticated drawing objects, or application-specific fields can open after conversion while still having lost the behaviour that made those items useful.

Braces, control words, and tables tell an RTF reader how to draw the document

A valid RTF document starts with a group such as {\rtf1. Braces create nested groups, and a backslash begins a control word or control symbol. For example, \b switches bold on, \i switches italic on, and \par ends a paragraph. The header can name the character set, default font, language, and document settings. Separate font and colour tables give later instructions a compact way to refer to a particular font or colour. This is structured text, not HTML: an unmatched brace, a missing header, or an instruction placed in the wrong group can make a reader reject the file or show part of it as literal code.

The RTF 1.5 generation introduced Unicode RTF and support for PNG, JPEG, and EMF picture types. Unicode characters can be represented with the \u control word plus a fallback-character count set by \uc. That detail matters during conversion. An RTF may declare an ANSI code page such as \ansicpg1252, yet also contain Unicode escapes. Treating the whole file as ordinary UTF-8 text and saving it back without preserving the RTF instructions can produce wrong quotation marks, accents, or non-Latin characters. Conversion software has to interpret the language, not merely replace the extension.

What an RTF handoff preserves, reduces, or deliberately leaves behind

• Readable formatting: plain paragraphs, basic character styling, and familiar page elements have standard RTF instructions, so they are the most portable portion of a document.
• Visible text rather than a sealed page: unlike a PDF, a normal RTF remains editable text; that is useful for revision but means its pagination can change on another computer.
• Font names, not guaranteed font availability: the font table can request a typeface, but it does not make that font appear on the receiving system unless font data was embedded and the reader supports it.
• Unicode with a conversion hazard: \u escapes can preserve characters beyond the declared code page, but a bad text-encoding round trip can make them display incorrectly.
• Older feature vocabulary: current Word-only objects may be converted to simpler content, pictures, or text because another RTF reader has no equivalent interactive object.
• No recovery by later up-conversion: converting the reduced RTF back to DOCX creates a new DOCX; it does not recreate discarded comments, controls, or original object data.

Use RTF as a compatibility copy when the required outcome is editable, basic rich text. Keep the source document if continued editing, exact design, accessibility structure, or live review features matter. If the requirement is a page that must look the same when printed, a verified PDF is usually the better delivery companion. The important choice is not whether RTF opens, but whether the target program must edit simple text or preserve a richer document model.

Current Word, LibreOffice, macOS TextEdit, and browser editors do not expose the same RTF feature set

Desktop Microsoft Word remains the most complete reference point for RTF 1.9.1 features because that specification was published for Word 2007. LibreOffice Writer opens and saves RTF, but its release notes continue to record RTF import fixes; LibreOffice 25.2.2, for example, lists a fix for automatic table padding on RTF import. That is useful evidence that an RTF file can be valid while particular layout details vary by program version. TextEdit can read and write rich text on macOS, but its simpler editing model is not a substitute for testing a complex Word document.

Web editors are a separate case. They often convert uploaded documents to their own internal model and may return DOCX, PDF, or an application document instead of maintaining every RTF instruction. A preview proves only that the previewer extracted enough content to display. For an editable delivery, open the output in the program actually named by the recipient and inspect fonts, page breaks, headers, tables, pictures, and any required field values. A test on the same operating system matters because missing fonts change character widths and can move text to a different page even when the RTF syntax is sound.

Wrong characters and “corrupt” warnings usually expose a header or encoding mismatch

One common failure is text that should contain an accent or curly quote appearing as unrelated symbols. In a documented encoding case, an RTF declared \ansicpg1252 but was written back as UTF-8 text; the reader followed the file’s ANSI declaration and decoded the bytes incorrectly. The fix is not to replace isolated bad characters by eye. Preserve the original bytes, inspect the initial RTF header, and use a converter that understands the declared code page and Unicode escapes. If the file is being moved as data, copy its bytes rather than reading it as plain text and writing a new text encoding.

Another real failure is a message that the RTF is damaged after it was generated or edited programmatically. RTF readers rely on balanced groups and a valid \rtf1 header. A font table that refers to the wrong font number, missing braces around a group, or an embedded-font block with bad length information can damage the remainder of the parse. Make a copy first, then open it in a plain-text editor only to inspect the beginning and obvious balance problem; do not casually delete backslashes or braces. If Word opens a copy using a repair route, compare the recovered result with the original visual reference because recovery may keep text while dropping formatting or objects.

If a table’s spacing changes only after opening in another suite, the problem can be a reader’s RTF import interpretation rather than corruption. Rebuild the affected table with simple cells and borders in the target program, or deliver an approved PDF when fixed layout is the true requirement. Resaving the file repeatedly through different applications is risky: each writer may replace unfamiliar controls with the subset it understands.

A compact RTF inspection table for deciding whether conversion is safe

Questions that matter before an RTF becomes the only copy

Is an RTF file just a TXT file with bold text?
No. It is a control language with nested groups, a header, formatting instructions, and optional tables for fonts and colours. It can be read as text, but its backslash instructions must be interpreted to display the document.

Why does an RTF look like code in Notepad?
Notepad shows the RTF instructions themselves, such as \b and \par. Open it in a rich-text editor to see the formatted result; do not remove those instructions if you need the formatting.

Why did text turn into question marks or strange characters?
The declared code page, UTF-8 bytes, and Unicode escape handling may no longer agree. Reconvert from the unchanged original rather than editing the damaged display character by character.

Can RTF guarantee identical pages on every computer?
No. It remains editable and can use substitute fonts, so line wrapping and page flow can change. Test the target program and use PDF for the fixed visual copy.

Should an RTF replace the original DOCX or ODT?
Not when the source contains advanced features or future-editing value. Retain the source and treat RTF as a tested compatibility derivative.