Convert RTF to PDF — the legacy format that survived Word, the web, and time

The short version

1. Got Word? Open the .rtf → File → Save As → PDF. Highest fidelity, handles embedded WMF images cleanly, exposes PDF/A for archival.
2. No Word? Drop the file into our converter. LibreOffice headless renders it server-side, same engine used by half the planet's archive workflows.
3. RTF with complex tables? Open in Word or LibreOffice first, visually check the table, then export. RTF table spec is fragile; silent-converters break nested rows.
4. RTF from a scientific instrument or legacy DOS tool? Character encoding is the trap — likely cp1252 or Mac Roman. LibreOffice auto-detects; some online converters guess wrong and produce gibberish.
5. Government form submission? Read the portal notes — they often want PDF/A-1, not plain PDF. Use Word or LibreOffice for that; most online converters don't emit PDF/A.
6. Legal filings, court docs? Many US courts accept RTF natively (PACER does). Convert only if the clerk asks for PDF.

What RTF actually is

Rich Text Format is a plain-ASCII markup language Microsoft published in 1987 so documents could move between Word, Mac MacWrite, WordPerfect, and every other word processor of the era. Open an .rtf in Notepad and you'll see text wrapped in curly braces and prefixed with backslash control words: {\rtf1\ansi {\fonttbl \f0 Times New Roman;} \f0 Hello world.}. That's a complete, valid RTF document.

The upside: every text editor on earth can read, diff, and edit an RTF. It survives email gateways that strip attachments. It can't carry a virus macro (unlike .doc, pre-2007). Its spec is publicly documented and stable.

The downsides compound in modern workflows: the table model is positional and fragile, image support is limited to WMF/EMF and a handful of raster formats, the style system is primitive compared to DOCX, there's no modern accessibility structure, and file sizes balloon compared to DOCX (RTF is verbose plain text; DOCX is a zipped XML bundle). Microsoft effectively froze the spec at RTF 1.9.1 (2008) and never updated it.

Why RTF still shows up in 2026

If you're asking "why do I have an .rtf in my inbox in 2026," the answer is almost always one of these:

• Legacy institutional systems. Courts, municipalities, some federal agencies, university records offices, and anything that standardized workflows in the 1995-2005 era. RTF was the "portable" format before PDF won.
• Scientific instruments and medical devices. Older lab printers, radiology systems, and medical records exports emit RTF because every system downstream can read it.
• Email-safe rich text. A few email systems (especially Outlook-to-Outlook on older Exchange) silently ship rich-formatted messages as RTF attachments named winmail.dat — the infamous TNEF payload.
• Cross-platform templates. Writers using Mac-only tools like Bean, Pages, or older Nisus Writer often export to RTF because every Windows recipient can open it.
• Government forms. State DMVs, IRS-adjacent third-party portals, VA, and several agencies still host RTF templates. You download, fill in, convert to PDF, submit.

How conversion actually works under the hood

RTF is a markup language; PDF is a fixed-layout drawing format. The converter has to parse the RTF tree, figure out where each line breaks at a given font size and page width, decide which font renders each run, flow tables, resolve embedded WMF/EMF vector images (which are themselves a Microsoft vector format nobody else implements well), and emit PDF drawing commands.

Where the paths diverge: Word runs its own high-fidelity layout engine — the reference implementation. LibreOffice has the deepest open-source RTF parser (it's been worked on since 1999). Pandoc treats RTF as semantic source and reflows it for its own LaTeX or HTML pipeline. Most cheap online converters lean on a stripped LibreOffice or a home-grown parser that handles 90% of documents and face-plants on the other 10%.

Path 1: Word's built-in Save as PDF

If Word is installed, open the .rtf (Word reads RTF natively, same File → Open dialog), then File → Save As → PDF. The layout engine is the reference implementation — you get pixel-identical output, clean WMF/EMF handling (because it's Microsoft's own vector format), and the PDF/A checkbox under Options.

The traps:

• Character encoding. Old RTFs sometimes declare \ansicpg1252 but contain Mac Roman or cp437 bytes — Word usually rescues these, but spot-check accented characters.
• Embedded fonts. Word embeds any font installed locally. If the RTF references "Garamond Pro" and your Word only has Garamond, substitution shifts line breaks.
• PDF/A for submissions. Check the ISO 19005-1 box under Options. Government portals silently reject plain PDF.

Path 2: LibreOffice (free, offline, deep RTF parser)

LibreOffice Writer has been hand-crafting its RTF parser since 1999. Open the .rtf, File → Export As → Export As PDF, pick your options. The dialog exposes PDF/A-1a, PDF/A-2b, page ranges, image compression, and tagged PDF for accessibility.

This is what we run on the back end. The advantage of doing it yourself: full control over the options dialog, offline-safe, no upload. The cost: ~300 MB install, a utilitarian UI, and the occasional layout drift from Word on very ornate tables.

Path 3: Online converter (no install, under 30 seconds)

Our RTF-to-PDF converter runs LibreOffice headless server-side — identical rendering to what you'd get from a desktop LibreOffice export, one drag-and-drop instead. No install, no sign-up, no watermark on the output. We accept .rtf, .doc, .docx, .odt through the same endpoint, so if your "RTF" is actually a .doc with the wrong extension (it happens), the tool still works.

What we won't do that Word does: proprietary WMF fonts Microsoft ships with Office (Marlett, MS Mincho, MS Gothic) need to be installed on the rendering server; we have the common ones, not all of them. For documents heavy in East-Asian text or uncommon Microsoft fonts, Word local is safer.

The table problem

RTF tables are the single biggest source of conversion defects. The spec describes tables positionally — each row is a sequence of cell-separator control words (\cell, \row) with column widths declared at the top. There's no proper nested-table model; complex RTFs simulate nesting with cell padding and baseline tricks.

What breaks: nested tables (three levels deep is brittle), cells with vertical-merge spans, auto-width columns that try to size to content, and tables that span page boundaries. If the output PDF has table rows collapsed or cells bleeding into margins, open the RTF in Word or LibreOffice first, re-save as DOCX, then convert. The DOCX table model is more robust.

Images: the WMF trap

RTFs often embed Windows Metafile (WMF) or Enhanced Metafile (EMF) images — Microsoft's vector formats. On Windows with Word, they render beautifully. On Linux/Mac through an online converter, the renderer has to translate WMF drawing commands to PDF drawing commands, and the translation is imperfect: gradients go solid, text inside WMF rasterizes blurry, dashed lines become solid.

Signals your RTF has WMF problems: small icons appear as white boxes, decorative flourishes render as solid black blocks, chart labels misalign. Workaround: open in Word or LibreOffice, right-click the image, export to PNG at 300 DPI, reinsert, save as DOCX, convert.

Encoding: the silent-corrupter

RTF predates Unicode mainstream adoption. Early RTFs are cp1252 (Windows Latin-1), Mac Roman, or cp437 (DOS). The file declares its encoding in the header: \ansicpg1252 means Windows Latin-1. Trouble starts when (a) the header is missing, (b) the header lies (byte content doesn't match declaration), or (c) the document mixes encodings mid-paragraph (rare but seen in RTFs joined from multiple sources).

Visible symptoms: accented characters turn into é or ’, the em-dash becomes a random punctuation glyph, curly quotes turn straight. Word and LibreOffice auto-detect; cheap online converters assume UTF-8 and produce gibberish. If your output looks like a mojibake parade, convert in LibreOffice with File → Open → select "Western (Windows-1252)" explicitly.

How the options compare (honestly)

Batch workflow — a folder of legacy RTFs

Got 200 court docs, insurance claims, or DMV templates to convert? Drop the whole folder into our converter. Global options (font embedding, image resolution) apply once to every file. Server-side LibreOffice renders in parallel; you get a ZIP with filenames preserved. CLI users can reproduce this with libreoffice --headless --convert-to pdf *.rtf — same engine, same result.

Works well / doesn't work

Works well

• Clean RTFs from Word, LibreOffice, Bean, or Nisus Writer
• Simple tables (rectangular, no vertical merges)
• Embedded JPEG/PNG images
• Standard cp1252, UTF-8, or properly-declared encodings
• Legal templates, court filings, plain-text-with-formatting docs
• Batches of 10-500 similar documents from legacy systems

Doesn't work (well) yet

• PDF/A archival submissions — use Word or LibreOffice
• Nested tables three levels deep — open in Word first, re-save
• Heavy WMF/EMF vector graphics — export images to PNG first
• Mixed-encoding RTFs from ancient concatenation workflows — open in LibreOffice with explicit encoding
• RTFs with East-Asian text requiring proprietary Microsoft fonts — use Word locally

Common questions

Is RTF the same as .doc?

No. RTF is plain ASCII markup readable in any text editor; .doc is Microsoft's binary Word format (1997-2003), and .docx is the zipped-XML format that replaced it in 2007. All three render similarly when opened in Word, but under the hood they're different specs. Our converter accepts all three.

Will my RTF's formatting survive the conversion?

Paragraph styles, bold/italic, font changes, headers, footers, simple tables, lists, and embedded raster images: yes. Nested tables, exotic typography, WMF vector graphics, and ornate multi-column layouts: sometimes. Spot-check pages 1, middle, and last before sending the PDF.

My RTF is from 2001 and has weird characters after conversion.

Encoding mismatch. The RTF claims cp1252 in its header, but the bytes are actually cp437 or Mac Roman. Fix: open in LibreOffice, File → Open, pick "Western (Windows-1252)" or "Western (Apple Roman)" explicitly in the encoding dialog, then re-export to PDF.

Can you preserve form fields from an RTF template?

Not as fillable fields in the output PDF — our converter (and Word's Save As PDF) produces static PDFs. To create fillable PDF forms, produce the static PDF first, then open it in Adobe Acrobat Pro, LibreOffice Draw, or PDFescape and add fields there. See our Word to PDF guide for the full form-PDF workflow — same logic applies to RTFs.

What about .rtfd from macOS?

.rtfd is Apple's "RTF Document Directory" — a macOS package (actually a folder) containing an RTF and its linked images. Right-click → Show Package Contents, extract theTXT.rtf inside, convert that. Or open in Apple TextEdit and Save As → PDF directly — TextEdit handles .rtfd natively.

Are my files private on your converter?

The RTF is uploaded for rendering (LibreOffice headless can't run in a browser tab), processed, and deleted within an hour. We don't log document content, train on files, or require an account. For maximum confidentiality (legal filings, medical records), Word's local Save As PDF never leaves your machine.

Going the other direction — PDF to RTF?

See our PDF to Word — it extracts editable text, which you can re-save as RTF from Word (File → Save As → .rtf). Clean PDFs with text layers convert well; scanned PDFs need OCR first.

Ready?

RTF to PDF →. Drop a .rtf, get a clean PDF with embedded fonts, preserved tables, and honest handling for embedded images. Free, no watermark, no sign-up. If the PDF comes out larger than you want, chain through our compressor. If you need PDF/A for a library or government submission, use Word or LibreOffice locally — we'll tell you so up front rather than silently emit a non-compliant PDF. And if you're also dealing with modern DOCX, read our Word to PDF guide — same engine, same rules. For more format-conversion head-to-heads, see our CloudConvert comparison.