Text to Speech

Cloud

Paste a text, pick an AI voice, download the audio as MP3.

Language
Voice
Style

Your text is encrypted in your browser and uploaded to our server. Only our server can read it. It is deleted as soon as it is spoken. The audio carries an inaudible watermark that marks it as AI-generated, as the EU AI Act requires. To learn more about how cut.audio handles your personal data, check our Privacy Policy. By pressing Generate speech you agree to our Terms of Service.

Listen to the Voices

AI voices. Each one reads the same short text, so you can compare them.

Narrator

Emma
Female voice
James
Male voice

News

Sarah
Female voice
John
Male voice

Audiobook

Kate
Female voice
Tom
Male voice

Radio

Lucy
Female voice
Mark
Male voice

Text to Voice in Three Steps

Type or paste your text, choose a language, a voice and a style, and press Generate speech. You hear the result right on the page and download it as MP3: a voice-over for a video, an intro for a podcast, or an article to listen to on the way.

An AI Voice Generator for Ten Languages

Each of the ten languages has a female and a male voice made for that language. Pick the language of your text and the voice that suits your project. There are four styles, narrator, news, audiobook and radio, and each has a female and a male voice of its own. Spanish is there for Spain and for Latin America.

From Text to Audio, Then Straight Into the Editor

The spoken file also lands in your audio pool. Open it in the editor to lay it over music, trim it in the cutter, or join several takes into one file.

Encrypted Before Upload
Only our server can read your text.
Ten Languages
English, Spanish, French, German, Chinese, Japanese and more.
MP3 Download
Download the audio or send it to the editor. WAV and FLAC with Pro.

Frequently Asked Questions

Yes. About 90 seconds of speech a day are free, one text can be up to a minute long, and you need no account. The daily allowance is shared with speech to text. Pro lifts these limits: up to 1.5 hours of speech a month instead of 90 seconds a day, one text up to 10 minutes long, and downloads as WAV and FLAC. That allowance is shared with speech to text as well.

English, Chinese, Japanese, Korean, German, French, Spanish, Portuguese, Italian and Russian. Each language has its own female and its own male voice.

Yes. MP3 is free; WAV and FLAC come with Pro. The audio is mono at 24 kHz.

Yes. Every file carries an inaudible watermark that identifies it as AI-generated speech. The EU AI Act requires AI-generated audio to be marked. Anyone can read it with a free, open-source detector: the steps are at cut.audio/check-ai-speech/.

Yes. The voice model needs a GPU, so the text is uploaded to our server. It is encrypted in your browser first, and only our server can read it. Your text is deleted as soon as it is spoken. The audio is deleted as soon as it has reached your browser, or after 30 minutes at the latest if you close the tab before that.

On our own GPU server in France. No third-party speech service is involved.