RTF is plain text made of groups in braces and control words with backslashes, which makes it look like a format you could convert with a regular expression. You cannot. Half the braces hold things that must never be printed, and the escape for an accented character is followed by a stand-in you have to count and discard.
Drop a file and the converted document appears before any PDF does, so the reading is visible. Then choose the page and convert.
Rich Text Format is nearly forty years old, was never properly retired, and still turns up constantly: as the output of an old system, as the format a court or a journal asks for, as the thing your word processor offers when nothing else will do. It is the lowest common denominator of word processing and it works, which is why it refuses to die.
What RTF is
Open an .rtf in a text editor and you can read it. That is the whole design: a document written as printable characters, so it survives systems that mangle anything else.
- Groups in bracesscope, and destinations
- Control words with a backslash\b, \par, \cell
- Escapes for anything else\'93 and \u8217
- Everything else is textwritten out plainly
Because it is text with no macros in it, RTF is one of the few document formats that is safe to open from a stranger, which is exactly why it became the format that systems exchange. It is also why it gets converted so often: it is easy to produce and inconvenient to read.
The braces that are not the document
A group in braces scopes formatting, so bold switched on inside one turns off again at the closing brace. Some groups do something else entirely: they are destinations, and their contents are not part of the document at all. The font table is one. So are the colour table, the style sheet, the document properties and the revision history.
This is the single most visible RTF conversion failure, and it is instantly recognisable: a document that begins with a list of typefaces. The fix is not a longer list of things to strip. It is reading the structure, so that anything inside a destination is skipped whatever it happens to contain.
The escape that eats the next character
RTF predates Unicode, and the way it was retrofitted is ingenious and hostile in equal measure. A character outside the code page is written as a control word giving its number, followed by a plain replacement for readers too old to understand the control word.
Miss that and the document is not broken enough to notice. Every accented letter simply has a question mark after it, every curly apostrophe too, and the file looks like it was typed by somebody in a hurry. It is the most common RTF bug there is, and it is why this counts them off properly and why the count is honoured when a file changes it mid-document.
What comes through
Headings, taken from the style sheet the file declares.
Bold, italic, underline, strikethrough, superscript and subscript.
Lists, with the marker group skipped so it does not become text.
Tables, cell by cell and row by row.
Alignment, accented characters, curly quotes and dashes.
Images, which are stored as hexadecimal and often as metafiles.
Footnotes, endnotes, running headers and footers.
Field codes, though any text they had already produced is kept.
Fonts and exact point sizes. One clean typeface, at the size you pick.
Merged table cells, which come out as separate cells.
Everything in the second column is reported when it is found, rather than dropped quietly. A converter that tells you what it could not do is worth more than one that produces a plausible file with a paragraph missing.
The rename that does not work
A surprising number of files called .rtf are not. Renaming a .doc changes the name and nothing else, and every tool that opens it then has to either guess or complain.
Telling them apart takes a text editor and a second. A real RTF begins with the characters {\rtf and is readable from there. A .doc begins with binary and looks like nonsense. A .docx is a zip, so it begins with PK.
Readable in any text editor.
So it is .docx or .odt.
Open and save as RTF first.
If yours turns out to be a Word file, the Word converter takes it directly and there is no need to save it as anything.
Where RTF still makes sense
- Handing a document to something old. Every word processor written since about 1990 reads RTF, which is not true of anything else.
- Getting text out of a system that only exports RTF. Case management software, medical records, legal databases and transcription tools all still do.
- Sending a formatted document that cannot carry a macro. RTF has no scripting in it, which makes it one of the few formats a cautious mail filter will let through.
- Not for keeping documents in. The files are large, the format is loosely specified, and two programs will disagree about the edges of it. Keep your documents in ODT or DOCX and export RTF when something needs it.
Why this runs in your browser
RTF is what old systems export, which means an RTF file is very often a record about a person: a case note, a discharge summary, a transcript, a letter from a solicitor. Those are exactly the files that should not be uploaded to a converter, and they do not need to be, because parsing text is something a browser does at speed.
The file is read in memory, laid out with real fonts, and checked by reopening the finished PDF before it is offered to you. Nothing is transmitted and nothing is stored.
The tool is right above
Drop an .rtf, look at how it read the document, and convert.
Common questions
The rename trap, the question marks, and what this deliberately leaves out.
01My file is called .rtf but it will not open. Why?
Almost certainly because it is not RTF. A real RTF file begins with the five characters {\rtf, and renaming a .doc or a .docx does not change what is inside it. Open it in any word processor and save it again as Rich Text Format.
02Will it keep my formatting?
Bold, italic, underline, strikethrough, superscript and subscript come through, as do headings, lists, tables and paragraph alignment. Fonts and exact sizes do not: the PDF is set in one clean typeface at the size you choose, because matching an arbitrary font list is not something a browser can do honestly.
03What about images in the file?
They are left out and reported. RTF stores pictures as hexadecimal text in the middle of the document, often in Windows metafile format, which is a drawing language rather than a picture. Pretending to handle that would be worse than saying so.
04Why did my accented characters come out clean when other converters leave question marks?
Because RTF writes an accented character as an escape followed by a plain stand-in for readers that cannot do Unicode, and the stand-in has to be counted off and thrown away. Converters that skip that step leave a question mark behind every accent in the document.
05Does it handle tables?
Yes. RTF marks the end of a cell and the end of a row with control words rather than closing tags, which is unusual but perfectly readable. Merged cells are not tracked, so a table with them comes out with the cells separate.
06Are footnotes and headers included?
No. Footnotes, endnotes, running headers and footers are separate destinations in the file and are skipped, with a note saying so. You can add a footer of your own here, including page numbers.
07Is RTF worth using at all now?
As an interchange format, sometimes: every word processor ever written reads it, it is plain text so it survives email and version control, and it carries no macros. As a format to keep documents in, no. Use ODT or DOCX and export RTF when something old needs it.
08Does the file leave my machine?
No. It is read and converted in your browser, and the PDF is checked by reopening it before you get it. Nothing is uploaded.