The only sure-fire way is to write your own RTF parser [spec], use a library like LibRTF, or you might consider keeping a RichTextControl open and updating it with new RTF documents rather than destroying the object every time.
I believe RTF is not a regular language, so cannot be properly parsed by RegEx (not unlike HTML, despite millions of attempts to do so), but you do not need to write a complete RTF parser.
I'd start with a simple string parser. Try:
- Remove content between {\ and }
- Remove tags. Tags begin with a backslash, \, and are followed by some text. If a backslash is followed by whitespace, it is not a tag.
- The document should end with at least one closing curly brace, }
Any content left which isn't whitespace should be document content, though this may have some exceptions so you'll want to test on numerous samples of RTF.