---
title: "How to convert text to speech and save an MP3 | ToolSugar"
description: "Turn typed or pasted text into spoken audio, choose a voice and language, and download an MP3. See which languages have a voice."
canonical: https://toolsugar.com/blog/how-to-convert-text-to-speech
image: https://toolsugar.com/og.png
---

# How to convert text to speech

Updated 2026-11-09

To convert text to speech, paste your words, pick a voice and download the audio. Text to Speech gives you an MP3 at 128 kbps. Every voice runs inside your browser, so the text stays on your device and is never uploaded.

[Open Text to Speech](https://toolsugar.com/tools/text-to-speech)

## Read your text aloud in three steps

Each job takes up to 3,000 characters. A counter under the text box shows how much you have used.

1. Pick a language, then pick a voice.
2. Paste or type your text. Spell out numbers if you want them read as whole words.
3. Press Make speech, listen to the result, and download the MP3.

## Which languages have a voice

A voice is a speech model your browser downloads and runs. The text never leaves your computer. English (US and UK), Spanish, French, German, Hindi, Italian and Brazilian Portuguese have voices.

Japanese, Chinese, Korean and Arabic have no voice yet, so the tool says so instead of making audio. Each run counts toward the 5 free AI runs shared across tools in any rolling 24 hours, with no sign up.

## Why the first listen takes a while

The first time you use a voice, your browser downloads its model. That is about 86 MB for most voices and 20 MB for German. Your browser keeps it afterward, so later runs only wait for the speech itself.

Speech is made one sentence at a time on your own processor. A long text takes longer, and a progress label shows which sentence is being read. You can cancel at any point.

## Get numbers and names read right

The voice reads what you type. A figure such as 1,250 may be read in a way you did not plan, so write it out in words when it matters. Play the result once before you share it, and fix any word that sounds wrong by respelling it.

## Questions

### Can I use a voice for Arabic?

Not yet, and the same goes for Japanese, Chinese and Korean. Pick one and the tool says so rather than producing silence.

### How long can the text be?

Up to 3,000 characters per job. For a longer piece, split it into parts and make one MP3 for each.

### Do I need an account?

No. Text to Speech shares 5 free AI runs with the other AI tools in any rolling 24 hours, and needs no sign up. The count uses a hash of browser traits and a hash of your IP.

[Open Text to Speech](https://toolsugar.com/tools/text-to-speech)

## Related guides

- [How to listen to a PDF](https://toolsugar.com/blog/how-to-listen-to-a-pdf)
- [How to turn text into a video with captions](https://toolsugar.com/blog/how-to-turn-text-into-a-video)
- [How to make a reel with AI from a topic](https://toolsugar.com/blog/how-to-make-a-reel-with-ai)
- [How to convert audio files to MP3, M4A or WAV](https://toolsugar.com/blog/how-to-convert-audio-files)

## Comparisons

- [ElevenLabs alternative](https://toolsugar.com/alternatives/elevenlabs-alternative)