Text to Speech

Type or paste text and download it as natural speech in Hindi, English and more.

Made in your browser — nothing is uploaded

About Text to Speech

Type or paste up to 5,000 characters and download them as natural speech in Hindi, English, Spanish, French, Italian or Portuguese — or, with an accent, in Bengali, Marathi, Tamil, Telugu, Urdu and seven more. Choose a male or female voice and the speaking speed. The voice is made in your browser, so your text is never uploaded.

How Text to Speech works

Text to Speech splits your text into sentences and reads each one with Kokoro, an open-source voice model (82 million parameters, Apache 2.0 licence) that runs in your browser. Before that, eSpeak NG turns each sentence into the sounds of its language, so Hindi words are pronounced as Hindi and English words as English. Short pauses are added after sentences and longer ones between paragraphs.

Hindi, English, Spanish, French, Italian and Portuguese have voices of their own. Bengali, Marathi, Gujarati, Punjabi, Tamil, Telugu, Kannada, Malayalam, Urdu, German, Russian and Indonesian are spoken by the closest available voice, so they have an accent. The voice model (about 90 MB) is downloaded once from the Hugging Face CDN and kept by your browser; your text is never sent anywhere.

When to use Text to Speech

  • Make a voice-over for a reel, short or explainer video without recording your own voice.
  • Listen to notes, articles or study material instead of reading them.
  • Create announcements, IVR greetings or shop messages in Hindi or English.
  • Check how a sentence sounds in a language you are learning.

How to use Text to Speech

  1. Type or paste your text. Up to 5,000 characters, in the script of its language — Hindi in Devanagari, for example.
  2. Choose language, voice and speed. Pick the language, then a male or female voice; slow it down to 0.6× or speed it up to 1.6×.
  3. Make and download. Click Make audio, listen to the result and download it as MP3 or WAV.

Options explained

Voice
English has eight voices, Hindi four (two female, two male); the other languages have one or two.
Speed
Changes how fast the voice speaks without making it higher or lower.
MP3 or WAV
MP3 is small and plays everywhere. WAV is uncompressed, for video editors and further editing.
Line breaks
Each new line becomes a short pause — use them to give lists and headings room to breathe.

Frequently asked questions

Can I use the audio in my YouTube videos and reels?

Yes. The voices come from Kokoro, an open-source model released under the Apache 2.0 licence, and the audio you make is yours to use, including in monetised videos.

Why does Hindi need to be typed in Devanagari?

The voice reads the letters it is given. Hindi written in Devanagari is pronounced correctly; Hindi typed in English letters (Hinglish) is not pronounced correctly.

Is there a character limit?

5,000 characters per run, about five minutes of speech. For longer text, make it in parts and join them with Merge Audio.

How are numbers read?

In English, numbers such as 1,250 or 12.5% are written out in words first. In the other languages the voice reads the digits in that language.

Does it work on a phone?

Yes, but the voice model is heavy: a computer makes the audio several times faster, and a phone needs a good connection for the first download.

Is Text to Speech free?

Yes. Text to Speech on DocEpic is free, with no sign-up, no watermark and no limit on how often you use it. The site is supported by ads.

Is what I type private?

Yes. Text to Speech runs entirely in your web browser, so what you type is never sent to us or stored anywhere.

Do I need to install any software?

No. Text to Speech works in any modern browser on Windows, macOS, Linux, Android and iPhone.

Related tools