Warren
English · Narration
For screenwriters and playwrights
Upload a script, give every character a voice, and choose whether a narrator reads the action or the scene plays as dialogue. Hear the pacing and the clunky lines before the real read-through.

Lewis Carroll
Alice's Adventures in Wonderland
Alice asks the grinning Cheshire Cat for directions — and gets none.
The Cat only grinned when it saw Alice. It looked good-natured, she thought; still it had very long claws and a great many teeth, so she felt it ought to be treated with respect.
Step-by-step guide
From an exported script to one file per act, with every speaker checked.
AI Table Read
What you get, from manuscript to finished audiobook.
Use cases
Where authors and storytellers put it to work.


Leads, character actors and a narrator for the action. Demo performances; in the studio you cast from standard and premium voice lists.
English · Narration
English · Narration
English · Narration
English · Narration
English · Narration
English · Narration
English · Narration
English · Narration
English · Narration
English · Narration
In depth
A table read is the first time a script leaves the page: actors around a table, reading it aloud from the first slugline to the last line. Writers run them because the ear is merciless. Dialogue nobody would say, a joke that needs one beat fewer, a scene that runs a page too long: all of it becomes obvious once it's spoken.
You can't book a room of actors for every draft. An AI table read fills the gap between drafts. FullCastBook gives every character a distinct voice, lets a narrator read the action or skips it, and hands you files you can listen to again. It won't give you an actor's interpretation, but it will show you where the script itself trips.
Turning a screenplay to audio starts with a file the studio can read: a PDF with a text layer, DOCX, TXT or pasted text. FullCastBook doesn't open Final Draft's .fdx files. Final Draft's own help pages describe saving a script as a PDF, or exporting it to other formats through File > Export Script on Mac or File > Export Document on Windows. Plain text keeps the line layout cleanest. Fountain scripts are already plain text, so rename the .fountain file to .txt.
PDF works, but its lines are re-flowed on import, so a character cue can end up at the start of its own dialogue line. If you have the choice, upload plain text or DOCX, and either way give the speakers a closer look.
The layout that gives the screenplay reader AI its clearest signal is the standard one: the character's name in capitals on its own line, the dialogue directly below. Stage plays follow the same idea, so the advice holds whether you need a screenplay or a play script reader. A short scene, written for this page:
Scene heading: INT. ALL-NIGHT DINER – NIGHT
Action: Rain on the window. Maya slides into the booth across from Leo.
Character cue, on its own line: MAYA. Dialogue on the next line: You said ten minutes.
Character cue: LEO. Parenthetical: (not looking up). Dialogue: I said ten-ish.
After casting, the scene heading and the action belong to the narrator, and each speech to its character. Speaker detection is AI-based and was built for novels, so check every scene: look for a line given to the wrong character, and for a cue name such as “MAYA” left behind as a narrator line. Delete those, or the narrator announces every speaker. Fix any line's speaker in a click.
A parenthetical is an acting note, and it works best as one. Take “(whispering)” out of the text and put it in the line's direction, where it becomes [whispering] and the voice performs it instead of reading it aloud. “(beat)” is a pause, not a word: delete it, or start the next line with a direction like [after a pause].
Extensions need the same tidy-up. “(V.O.)” and “(O.S.)” can go, since the voice already belongs to the character; “(CONT'D)” should go everywhere, or you'll hear it every time a speech breaks across action. For more on what directions and sound tags like [sigh] can do, see multiple voices text to speech.
The old rule of thumb says a screenplay page is about a minute of screen time, so a 100-page feature reads in roughly 100 minutes. A dialogue-only table read runs shorter, because action that would take screen time takes no time at all. Use the read for relative pacing, not a stopwatch: which scenes drag, which exchanges rush, and where an act lands flat.
Listen for the things a script reader AI catches better than your eyes: lines that are on the nose, two characters who sound alike on the page, speeches nobody could say in one breath. Rewrite in the studio and regenerate just those lines; regenerating uses credits again, so batch your fixes.
Every chapter exports on its own, so if each act or sequence is a chapter, you get one MP3 per act, a ZIP of them all, or an M4B with a marker at every act break. Send them to a co-writer, a producer or your writing group, say it's an AI read, and collect timestamped notes. Standard voices are fine for a notes pass; see the plans for longer scripts.
An AI table read is a listening draft: plain voices, no music, no sound design, made to hear the script and fix it. An audio drama is a finished production, with every scene performed, mixed and released. If the script is meant for the ear, the same project can grow into one: see the AI audio drama generator for how scene headings, sound cues and music are handled, and hear what a directed cast sounds like in the examples on the home page.
A read-through of your script by AI voices instead of actors: each character speaks in a distinct voice, and a narrator can read the action. It's a quick way to hear pacing and dialogue between drafts, not a replacement for actors.
Yes. FullCastBook detects who speaks each line and gives every character a voice. Detection is AI-based, so check and fix the speakers before you generate.
No. Export the script from Final Draft as a PDF or plain text and upload that. Plain text keeps a character's name on its own line, which is the layout to aim for.
Yes. Sluglines, action and stage directions arrive as narrator lines; delete the ones you don't want heard, or remove them from a copy of the script before you upload.
As a rule of thumb, a screenplay page runs about a minute, so a 100-page feature reads in roughly 100 minutes, and less if you skip the action. Credits depend on the characters voiced; the audiobook cost calculator estimates them.
Yes. Export one MP3 per act or scene, or a ZIP, and send the files. They're yours under our terms; just say it's an AI read when you share it.
Upload a screenplay or stage play, check the speakers and hear your first scenes free. No card needed.