Can I download the speech as an MP3?

Yes, for English, German, French, Spanish, Italian, Ukrainian and Japanese text: choose a studio voice, then save the result as MP3 or WAV. Your device's built-in voices read every other language aloud, but they play live and usually cannot be saved as a file.

Which languages can it read?

Built-in voices cover whatever languages your device has installed (Windows, macOS, Android and iOS each ship different ones). The downloadable studio voices speak English, German, French, Spanish, Italian, Ukrainian and Japanese.

Is there a character or usage limit?

There is no account, no API key and no daily quota. Long texts are split into chunks and read one after another; one studio-voice file holds up to 50,000 characters (20,000 on a phone).

Yes. Every tool is free to use, with no account and no credit card. The work runs on your device rather than on our servers, which is what keeps it free to run.

Does it work on mobile?

Yes. It runs in current mobile browsers on iOS and Android, with nothing to install.

🔊

Free Text to Speech

Read text aloud free with your device's voices. English, German, French, Spanish, Italian, Ukrainian and Japanese text can also be saved as MP3 or WAV with a studio voice. No API key, no upload, no cost.

Runs in your browserNo sign-up, no pop-upsNo login requiredFree forever

Many voice sites meter free use in characters or credits — here the voices run on your device, with no account.

Text to speak

Import file

106 characters · ~8s spoken

Voice type

Voice (19 available)

Rate (1.0x)

More settings

Pitch (1.0)

Volume (100%)

Device voices can't be saved. Need a file?

🔒 Your text stays on this device. Voices come from your browser or operating system, and nothing is uploaded.

Common uses

Proofread by ListeningHave your writing read aloud to catch typos, awkward phrases, and missing words your eyes skip over.
Make Long Documents AccessibleHave articles, PDFs, or emails read aloud for users with dyslexia, vision impairment, or reading fatigue.
Pronunciation Practice for Language LearnersHear words and phrases pronounced in any of dozens of languages.
Preview Audio for Content CreatorsHear how your script will sound before recording, or generate quick voiceovers for prototypes.
Listen While You MultitaskHave articles, documentation, or emails read aloud while you exercise, commute, or do chores.

How It Works

Two kinds of voice, both running on your device: your system's built-in voices, which play instantly, and a studio voice that can be saved as an MP3 or WAV file. No API key and no account.

1
Choose device or studio voice
Device voices come from your operating system and speak many languages, but browsers do not let a page record them. The studio voice speaks English and produces a file you can download.
2
Your text is turned into speech on your device
A device voice is spoken by your operating system's speech engine. The studio voice downloads a small AI model and a pronunciation engine once (about 105 MB together, from LibraTool's own servers) and generates the audio in your browser.
3
Listen, then save
Adjust speed and volume and play it back. With the studio voice, download the result as an MP3 or WAV file.

Your text is never sent to LibraTool. Studio-voice audio is generated in your browser. Device voices are provided by your system; some browsers (notably Chrome on Android) may use an online voice service for certain voices.

Guide

Worked example

A 1,500-character article (about 250 words) read by a studio voice comes to roughly a minute and a half to two minutes of audio. Saved as WAV that is about 5 MB (24 kHz mono 16-bit), and MP3 is much smaller. The first time, the English voice downloads about 105 MB (the voice model plus a pronunciation engine) and is then cached. The same text read by a device voice plays at once but cannot be saved as a file.

Which settings to use, and the limits

Pick the voice type by what you need. "Device voice" plays instantly in any language your system has installed, but browsers do not let a page capture it. "Studio voice" is a one-time download that works offline afterwards, and it is the only type that produces a file.
Know which languages can be saved. Studio voices exist for English (American and British), German, French, Spanish, Italian, Ukrainian and Japanese. The first download is about 105 MB for English, 96 MB for German or French, 47 MB for Spanish, 82 MB for Italian, 77 MB for Ukrainian and 105 MB for Japanese. A studio voice cannot read text in another alphabet.
Rate is 0.5x to 2x; the studio voice has no pitch. The rate slider moves in steps of 0.1. On a studio voice, speed is applied while the audio is created, so changing it re-creates the file. Pitch applies to device voices only.
Long texts are split, and you can stop early. One studio file holds up to 50,000 characters (roughly an hour of speech) on a computer and 20,000 on a phone. Longer text is split into numbered files. The text is made in pieces of about 300 characters, and "Stop and keep" saves what is done so far.
WAV is ready at once; MP3 needs an encoder. WAV needs nothing extra. MP3 and M4A download a one-time audio encoder of about 30 MB. On mobile data, the tool asks before starting a large first download.
Import a document instead of pasting. Drop a PDF, Word (.docx), EPUB, TXT or HTML file onto the text box. Files up to 200 MB are read on your device. A scanned PDF holds only pictures of pages, so run it through the OCR tool first.

Common problems and how to fix them

A device voice says it cannot read the text.: That voice does not support the language of the text. Pick a voice for that language in the list, or install one in your system's speech settings and reload the page.
There is no voice for my language, or no voices at all.: Device voices come from your operating system, so what is available differs on Windows, macOS, Android and iOS. Add a voice in the system speech settings and reload. For English, German, French, Spanish, Italian, Ukrainian or Japanese you can use a studio voice instead, with nothing to install.
The studio voice says it cannot read this alphabet.: Each studio voice reads one language. The English voice cannot read, for example, Cyrillic text. Choose the studio voice for the text's language, or a device voice. This check runs before any download, so nothing is fetched.
The text is too long for the studio voice.: The limit is 50,000 characters on a computer and 20,000 on a phone, and the check runs before the model is downloaded. Shorten the text, split it into parts, or use a device voice if you only need to listen.
My PDF imported with no text.: A scanned PDF is made of page images, not text. Run it through the OCR tool, then paste the result. A password-protected or damaged PDF also cannot be read.

How it compares

A device voice starts at once and covers many languages, but the audio plays through your speakers and cannot be saved by the page. A studio voice is slower to produce, needs a download of roughly 80 to 105 MB the first time, and covers only three languages, but it gives you a WAV or MP3 file you can keep and reuse offline.