How to Convert a PDF to FDX

A PDF with a real text layer can be rebuilt into an editable Final Draft file in the browser, one element at a time. A scanned PDF cannot — it is a photograph of a script, and without OCR there is no text in the file to recover. Either way the conversion is a rebuild rather than a restore, so pagination, dual dialogue and page-break markers will not come back intact.

Updated 2026-09-18 · About 6 sections · 5 answered questions

When you end up converting a PDF back to FDX

Three situations produce this search, and they have different outcomes:

You have your own script as a PDF only. You exported a PDF, the laptop died, or you wrote the draft in software you no longer have. This is the good case: the PDF was born from a screenplay document, so the text and the element positions are intact.
Someone sent you a PDF draft. A producer, a manager or a collaborator sends a locked PDF of a draft. If you are the writer and you lost the source, converting is the only way back to an editable file. If you are not the writer, remember that a PDF sent for reading is not an invitation to rewrite it.
The script exists only on paper. A printed or scanned script — a draft from 1998, a competition copy, a photocopy. There is no text in that file at all, and no converter can invent it.

The first case converts well. The second converts well if the PDF is text-based. The third needs OCR before any of this is possible.

Why a scanned PDF cannot be converted without OCR

A PDF does not contain a script. It contains a description of where to print ink on a page. There are two very different kinds of file behind the same .pdf extension:

Text-based PDF: Produced by exporting from screenwriting software or a word processor. Every letter is stored as a character at a known position. Selecting a line of dialogue highlights that line.
Scanned or image PDF: Produced by a scanner, a phone camera, or a photocopier. The file contains a picture of the page and nothing else. Selecting text highlights a rectangle of image, or nothing at all.

You can tell which one you have in ten seconds: open the PDF, click at the start of a line of dialogue and drag across it. If the text highlights, there is a text layer and you can convert it. If the whole page highlights as one block, or nothing does, the script is a picture.

OCR — optical character recognition — is what turns that picture back into characters, and it is a different job from conversion. It is also a worse job, for two reasons. First, OCR on Courier misreads the glyphs that monospaced type makes ambiguous: 1, l and I; 0 and O; rn rendered as m. Character and scene numbers are usually the casualties. Second, and more damaging, screenplay conversion depends on geometry: a line that starts at the character-cue column is a character name, a line indented 1 inch past the action margin is dialogue. OCR gives you lines of text with rough positions, and rough is not enough when the difference between action and dialogue is a third of an inch. Expect to re-tag most elements by hand on an OCR'd feature, which is several hours of work — the reason most people treat a scan as a retyping job and start clean.

Converting a text-based PDF to FDX in the browser

The free PDF to FDX Converter reads a screenplay PDF in the browser and returns a Final Draft XML file. Drop the PDF, it recovers the text and maps each line to a screenplay element, and you download the .fdx. Nothing to install, no account.

Open /tools/pdf-to-fdx/.
Drop the PDF in. It reads the text layer and measures each line against the standard margins — action at 1.5", dialogue at 2.5", character cues near 3.7", transitions right-aligned — plus the all-caps and blank-line patterns that mark scene headings and character names.
Read the preview. The tool shows what it recovered and how it typed each element. This is the moment to notice a problem: if the preview shows page after page of action, your PDF is either scanned or laid out in a font and margins that are not screenplay standard.
Download the FDX. You now have a file you can open in Final Draft, Fade In, WriterDuet or Zen Writer and edit.
Do the check pass before you write a word of new draft. The last section of this article is the checklist, and it is not optional — a mis-tagged element that survives into your next draft costs more to find later.

One practical note: the tool's page states that uploaded PDFs are stored by Zen Writer. For a script under an NDA that may matter, in which case export the FDX from desktop software instead.

Why the conversion is approximate: a PDF has no screenplay elements

This is the part most conversion advice skips, and it explains every item in the table below. Final Draft XML stores meaning: this paragraph is a character cue, that one is dialogue, this scene heading carries the number 7. A PDF stores appearance. Nothing in a PDF says "scene heading" — the words are just glyphs sitting at x-position 108.

Walk the same three lines through both formats and the difference is obvious. In the PDF, the page is only a list of positioned text:

INT. WAREHOUSE - NIGHT

Dust hangs in the light. MARCUS steps over a cable.

MARCUS
You said the doors would be open.

In the FDX, each line carries its type, and the app decides where it sits:

<Paragraph Type="Scene Heading">INT. WAREHOUSE - NIGHT</Paragraph>
<Paragraph Type="Action">Dust hangs in the light...</Paragraph>
<Paragraph Type="Character">MARCUS</Paragraph>
<Paragraph Type="Dialogue">You said the doors would be open.</Paragraph>

Converting means inferring the second from the first. The inference is reliable when the PDF was generated from a standard screenplay template, and it degrades in predictable ways: an A4 page instead of US Letter shifts the margins slightly, so the columns do not land where the detector expects; a PDF exported from Word in Times New Roman is not in screenplay geometry at all, and leads to a script of nothing but action paragraphs; a PDF that has been re-flowed, compressed or re-exported several times has letter positions that no longer line up column by column. In those cases you get a text dump rather than a formatted script, and the honest move is to accept it as text and re-tag it in a real editor.

What converts cleanly and what needs a manual pass

This is the realistic scorecard for a standard spec-format PDF exporting to FDX. "Manual pass" does not mean retyping — it means reading through and fixing the element or the metadata.

Element	Result	Notes
Scene headings	Clean	Detected from `INT.`/`EXT.` and the all-caps line before a blank line.
Action	Clean	Reflowed to the action margin.
Character cues	Clean	An all-caps line sitting between a blank line and indented text.
Dialogue	Clean	Contiguous indented text under a cue.
Parentheticals	Clean	Detected from the surrounding parentheses and the cue above.
Transitions	Usually clean	A right-aligned line ending in `TO:`. Unusual right margins can demote them to action.
Title page	Manual pass	The text survives; the centring, the position on the page and the contact block are re-laid out.
Page breaks and page numbers	Lost	Page numbers are print artefacts, not content. The new file paginates to its own layout, so page 24 of the PDF will not be page 24 after conversion.
`(MORE)` and `(CONT'D)`	Manual pass	Page-break markers from the old layout. Read through and remove the orphans.
Dual dialogue	Lost	Two side-by-side speeches come back as two consecutive dialogue blocks, because the PDF stores them as two separate text runs at two sets of coordinates. Re-flag them by hand.
Scene numbers	Lost	If the PDF was a shooting script, the numbers in the margins are stray text, not attributes on the scene.
Revision colours	Lost	Colour is not in the text layer. Every revised page looks like every other page.
Bold, italic, underline	Unreliable	Depends on whether the PDF kept the font variants rather than substituting one font.
Non-standard fonts and margins	Risky	Element detection is geometric. Proportional fonts and A4 pages move the columns and mis-tag elements.

How to check the converted script

Ten minutes here saves an afternoon later. Work through this list on every conversion:

Compare the last page number. A converted feature should land in the same range as the PDF — usually within a page or two. A 108-page PDF that becomes 92 pages means you have lost elements, most often merges of dialogue, and you should look for where.
Count the scene headings. Count them in the PDF and in the new file. A missing slugline usually means one got typed as action or fused into the line above it.
Read the top and bottom line of each page. That is where dialogue split across pages, orphaned character cues and stray (MORE) markers live.
Find the dual dialogue. Search for the two scenes you know use it and rebuild them. Nothing in the converted file flags them.
Check every transition. CUT TO:, SMASH CUT TO:, MATCH CUT TO: should be right-aligned transition elements, not action lines, and should not be followed by a blank page of nothing.
Check the title page. Rebuild it rather than trusting the conversion — it is a page of centred text and every conversion mangles the position. The Title Page Generator handles the layout conventions.
Retype the first and last scene. Not for accuracy so much as calibration: if those two read correctly, the middle almost certainly does.

After that, save the converted file as a new draft name rather than overwriting anything — Title_v2_from-PDF.fdx — so the original PDF stays the reference copy if a tagging problem turns out to be systemic.

Frequently asked questions

Can any PDF be converted to FDX?

Only a PDF with a text layer. If the script was scanned or photocopied, the file is an image and there are no characters to recover until OCR has run. You can test the file in five seconds by trying to select a line of dialogue.

Why is my converted script a different number of pages?

Because the page breaks were part of the PDF's layout, not part of the script. The new FDX paginates to its own template, so its page 30 is not the PDF's page 30. The total count should land close to the original; a large drop usually means dialogue was merged.

Does converting a PDF to FDX keep dual dialogue?

No. Side-by-side speeches are stored as two separate blocks of text at two sets of coordinates, and a converter has no reliable way to know they are meant to sit side by side. They come back as two consecutive speeches and must be re-flagged by hand.

The PDF says it is locked. Can I still convert it?

Often yes. Most producer PDFs are only restricted from editing and printing, and text extraction still works. If the file needs a password just to open, you need that password; if the owner has blocked copying entirely, the converter will return little or nothing.

Does the FDX I download still open in Final Draft?

Yes. FDX is Final Draft's own XML format and opens in Final Draft as well as compatible apps including Fade In, WriterDuet and Zen Writer. Note that a different app may paginate the file slightly differently again, so re-check the page count after importing.

Ready to turn this idea into a screenplay?

Zen Writer opens the converted FDX and reformats mis-tagged elements as you fix them, so cleaning up a PDF conversion takes minutes instead of an afternoon.

Open the Converted FDX in Zen Writer

Free on iPhone, iPad, Mac and Android.

Continue reading

FDX & Final Draft Files

How to Convert a PDF to FDX

When you end up converting a PDF back to FDX

Why a scanned PDF cannot be converted without OCR

Converting a text-based PDF to FDX in the browser

Why the conversion is approximate: a PDF has no screenplay elements

What converts cleanly and what needs a manual pass

How to check the converted script

Frequently asked questions

Ready to turn this idea into a screenplay?

Continue reading

What Is an FDX File?

How to Open an FDX File (With or Without Final Draft)

How to Convert FDX to PDF

FDX vs Fountain: Which Screenplay Format Should You Use?

Screenplay Formatting Examples:

PDF to FDX Converter

Industry Title Page Generator

More in fdx & final draft files