TutuSound

Turn text into someone actually telling it

Multi-voice casting · ambience beds · sound effects and score — a whole book in one run

Every character

Figures out who is speaking; voices never blur together

Three-layer mix

Ambience + effects + score — not a flat read-aloud

Batch factory

505 episodes run to completion with QC; closing the tab loses nothing

Real run: 505 episodes · 1.23M characters · 78.5 hours · every episode verified on five axes

More: what it does · FAQ

Free online text to speech — audiobooks, multi-voice narration, voice cloning

TutuSound is an audiobook generation platform that runs in your browser. Paste some text and hear it spoken; upload a whole novel and it splits the book into episodes, works out who says each line, assigns a voice to every character, and exports audio you can publish as is.

The interface is available in 16 languages. Your text is stored as a structured project, so every step can be undone. The AI assistant can run the whole pipeline for you — splitting, casting, auditioning, mixing, QC, export — or you can do each step by hand. The platform ships no content of its own; it only processes text you upload and have the rights to.

What it does

Text to speech

Paste text, pick a voice, hear the result in seconds.

Audiobook production

Upload a whole book, get it split into episodes and generated in one pass.

Multi-voice narration

Detects who speaks each line and keeps one voice per character.

Voice cloning

Give it a short reference clip and it reads anything in that voice.

Chapter splitting

Splits by length and meaning; merge or split any boundary by hand.

Ambience and effects

Three-layer mix of voice, ambience and effects with automatic levelling.

QC and retakes

Checks duration, channels, silence and pace per episode; retake just the bad ones.

Batch export

Finished audio plus metadata, packaged by volume, ready to publish.

AI assistant

Run the whole pipeline from one sentence, and watch every step it takes.

Text to speech
Open the workspace

Frequently asked questions

Which text formats are supported?
txt, md, epub, and plain paste. Upload a whole book at once — it finds the chapter boundaries itself, and you can merge or split any of them afterwards.
How long does a book take?
It depends on length and the current queue. A 500,000-character novel usually takes a few hours. You can close the page while it runs and check the workspace when it is done.
Can I choose the voice?
Yes. Pick from the built-in voices, or clone a new one from a short reference clip. In a multi-character book you can set a voice per character.
Who owns the generated audio?
You do. The platform ships no content of its own and only processes text you upload. Please make sure you hold the rights to whatever you upload.
How many languages are supported?
The interface is available in 16 languages. Speech synthesis currently covers Chinese and English, with more being added.
Does it keep going if I close the browser?
Once the job is handed to the cloud, yes — the whole pipeline runs on the server and closing the page changes nothing. Until it is handed over, the page tells you plainly not to close it.