Multiple Voices Text to Speech, Assigned Automatically

Free credits on sign-up. No card needed.

Paste a dialogue, a script or a short scene. FullCastBook works out who says each line, gives every speaker a different AI voice and turns the whole conversation into one audio file.

Try it on your scriptHear the demo
TextM4BDemo
txt

my-story.txt

Text

Reading your file…

Alice's Adventures in Wonderland — The Cheshire Cat
NarratorNaomiAliceAbbyCheshire CatSnik How it works
Alice's Adventures in Wonderland cover

Lewis Carroll

Alice's Adventures in Wonderland

Alice asks the grinning Cheshire Cat for directions — and gets none.

NarratorNaomiStoryteller

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.

AliceAbbyCurious

Cheshire Puss, would you tell me, please, which way I ought to go from here?

Cheshire CatSnikSly

That depends a good deal on where you want to get to.

AliceAbbyCurious

I don't much care where—

Cheshire CatSnikSly

Then it doesn't matter which way you go.

AliceAbbyCurious

—so long as I get somewhere.

Cheshire CatSnikSly

Oh, you're sure to do that, if you only walk long enough.

~1 min · 3 voices · ready to play7
Free preview, no signup to listen. Your own chapters & export need a free account.I'm ready →
AI voices to cast
137
70 standard · 67 premium
file types in
5
EPUB · PDF · DOCX · TXT · MD
ways out
3
M4B with chapters · MP3 · ZIP
voice models
4
Gemini · ElevenLabs

Step-by-step guide

How to use Multi-Voice Text to Speech

Free to start: no card, nothing to install.

  1. Paste your lines

    Paste a scene, a script or a conversation, or upload a TXT, DOCX, PDF or EPUB file. Plain prose works, and so does a script with a name before each line.

  2. Check who says what

    Every line is assigned to a speaker, or to the narrator for prose between the lines. Change the speaker of any line before you generate.

  3. Give each speaker a voice

    Pick a different voice for every speaker, then add direction in brackets, such as [nervous, speaking fast], or sound tags like [laugh] where a line needs them.

  4. Listen and download

    Play it in the browser, regenerate any line that lands wrong, and download one MP3 of the whole conversation.

Multi-Voice Text to Speech

Features & capabilities

What you get, from manuscript to finished audiobook.

  • Speakers found for you

    No speaker IDs to type. Each line is matched to the person who says it, from the names in front of lines, dialogue tags and the flow of the conversation.

  • A different voice per speaker

    Every person in the conversation gets a voice of their own, with a separate narrator for any prose between the lines.

  • Direction in brackets

    Write how a line should sound, like [whisper, like sharing a secret], and add sound tags such as [laugh], [sigh] or [breathe].

  • Fix one line, not the take

    Regenerate a single line after an edit or a recast instead of rendering the whole dialogue again.

  • Many languages

    Voices read many languages, with English working best today, so you can build dialogues for learners as well as listeners.

  • One MP3, or one per chapter

    Download the conversation as a single MP3. Longer pieces can also go out as an M4B with chapter markers or a file per chapter.

Use cases

What can you make with Multi-Voice Text to Speech?

Where authors and storytellers put it to work.

  • Audio fiction

    Turn a short story or a single scene into a cast recording, with a narrator for the prose and a voice for each character.

  • Explainer dialogues

    Teach an idea as a conversation between a curious learner and an expert, which is often easier to follow than one voice lecturing.

  • Language practice

    Write a café or airport dialogue, give each role a distinct voice, and let learners hear who is speaking as they follow the text.

  • Script table reads

    Hear a screenplay or a stage scene read by a cast before rehearsal, to catch clunky lines and slow pacing early. More in AI table read.

Hear example character voices

Demo performances. In the studio you cast from standard and premium voice lists.

More examples
On this page
  1. 1Multiple voices text to speech in one pass
  2. 2Multi speaker text to speech: how speakers are detected
  3. 3Why do-it-yourself setups stop at two voices
  4. 4Dialogue text to speech for scripts and scenes
  5. 5Conversation text to speech with [laugh] and [whisper]
  6. 6Use cases: audio fiction, explainers, language practice
  7. 7Free credits and MP3 export

In depth

Multiple voices text to speech in one pass

Multiple voices text to speech reads a conversation the way people hear one: each speaker has their own voice, so you never need "she said" to know who is talking. It suits anything built on back-and-forth, from a scene in a story to a two-person explainer, a script or a language lesson.

Most text-to-speech tools read everything in one voice, and the ones that do more usually make you mark every line by hand. FullCastBook takes the whole text, finds the speakers, gives each one a voice and returns a single audio file. You review the cast, adjust what you like and download.

Multi speaker text to speech: how speakers are detected

Multi speaker text to speech only works if every line goes to the right person. Before any audio is made, FullCastBook reads the text and assigns each line to a speaker, using the names in front of lines, dialogue tags such as "asked Maya" and the turn-taking of the conversation. In prose, everything outside the quotation marks goes to the narrator.

You don't need special markup. A script with ALICE: and BOB: in front of each line is the easy case, and plain prose works too. Detection is a first pass, so skim any stretch where three people talk without tags and change a speaker that's wrong. It takes seconds, and the rest of the scene follows your cast.

Why do-it-yourself setups stop at two voices

If you have tried to build multiple voices text to speech yourself, you have met the limits of the APIs. As of October 2026, Google's Gemini TTS handles up to two speakers in a single request and suggests generating each turn separately for more, while ElevenLabs' Text to Dialogue has no speaker limit but recommends keeping each request to about 2,000 characters.

That is fine for a short exchange. A ten-minute scene with four people means splitting the text, generating it piece by piece and stitching the clips back together in the right order. FullCastBook does that bookkeeping for you: each line is voiced by its speaker's voice and the result comes back as one continuous file.

Dialogue text to speech for scripts and scenes

Dialogue text to speech lives or dies on contrast. Two voices of the same age and pitch blur together after a few lines, so choose voices that differ clearly: an older and a younger one, a low and a bright one, two different accents. If the scene has a narrator, keep that voice calmer than the characters.

Format matters less than you might think. A line of description between two quotes is read by the narrator, and stage directions in parentheses can stay as narration or be cut before you generate. Long speeches read best split into paragraphs, so the voice has natural places to breathe.

Conversation text to speech with [laugh] and [whisper]

Real conversations are full of small noises, and conversation text to speech sounds flat without them. Add sound tags such as [laugh], [sigh] and [breathe] where they belong, and put a written direction in brackets before a line, like [whisper, like sharing a secret] or [annoyed, cutting in]. Pauses can be directed the same way: [after a long pause, quietly].

A few well-placed notes do more than many. Mark the turning points: the joke that lands, the moment someone gives in, the line that should be barely audible. If a delivery still misses, adjust the note and regenerate that one line. More on writing these notes is in AI voice emotion tags.

Use cases: audio fiction, explainers, language practice

Writers use multiple voices text to speech to hear a scene before committing to a full audiobook. Teachers build listening exercises where each role sounds different, and creators turn a two-person explainer into something closer to a podcast.

When the text grows into a whole book, the same studio makes a multi-voice audiobook with chapters, a narrator and a cast that stays the same from book to book. For spoken stories and read-alouds, see AI voice for storytelling.

Free credits and MP3 export

Trying multiple voices text to speech costs nothing up front: sign up and you get free credits with no card, enough to hear a scene in several voices. Credits are counted by the characters that get voiced, so a short dialogue costs little, and regenerating a line uses credits again. When you need more, compare the plans.

Export one MP3 of the whole conversation, or, for longer projects, an M4B with chapter markers or a ZIP with one MP3 per chapter. The files are yours. You can hear example performances on the home page, and when you're ready for a book, start in the AI audiobook generator, where you cast from the standard and premium voice lists.

FAQ

Multi-Voice Text to Speech: FAQ

Something not covered here? Email us at support@fullcastbook.com

Can text to speech use two voices at once?

Yes, and more than two. Multiple voices text to speech gives each speaker in your text their own voice and puts the lines in order in one file. The voices take turns, as in a real conversation, rather than talking over each other.

How many speakers can one script have?

In multiple voices text to speech, every speaker the detection finds gets a place in the cast with a voice of their own. Listeners follow a scene most easily when each voice is clearly different, so give the main speakers your most distinct voices.

Do I need to label each speaker?

No. Names in front of lines help, but plain prose with dialogue tags works as well. You can change the speaker of any line before generating.

Can I add pauses and vocal sounds?

Yes. Add sound tags such as [laugh], [sigh] or [breathe], and write delivery notes in brackets, such as [after a long pause, quietly]. If a line still misses, tweak the note and regenerate just that line.

Is multi-voice text to speech free to try?

Yes. New accounts get free credits with no card, enough to hear a scene in several voices. Longer projects need a paid plan; see pricing.

Can I use the audio commercially?

You own the files you export, and commercial use follows our terms of service. If you voice someone else's text, make sure you have the rights to it.

More in this collection

AI audiobook generator: start hereUpload a manuscript, give every character a voice and export MP3 or M4B with chapters.

Give every speaker a voice.

Paste a dialogue, cast the voices and download one MP3. Free to start.

Try it freeSee pricing