Private vs cloud text to speech

Private vs cloud text to speech: how they differ.

Some text to speech runs on your own device; some runs on a company’s servers. The difference decides where your script goes, which voices and languages you get, how fast it is and how you pay. KUKU Labs offers both, so here is how to choose.

In short: Private text to speech makes the audio in your browser, so your text isn’t uploaded; KUKU Labs has 102 private voices in five languages, unlimited on Basic. Cloud text to speech sends the text to a speech provider’s servers; KUKU Labs has 1,108 premium cloud voices in 48 languages, with a monthly character allowance.

How it works

  1. 1

    Decide what the script needs

    Private and unlimited for English, French, German, Spanish or Portuguese; premium cloud voices for 48 languages, including Urdu, Hindi and Arabic.

  2. 2

    Hear both

    Every private and premium voice has a free sample. Listen with the same kind of script you will use.

  3. 3

    Generate and download

    Pick the tool in the Studio, generate, and download MP3 or WAV. You choose private or cloud every time.

Private and cloud text to speech side by side

How KUKU Labs’ private voices and premium cloud voices compare
Private voicesPremium cloud voices
Where the audio is madeIn your browser, on your own deviceOn our cloud speech provider’s servers
Does your text leave your device?No: it isn’t uploadedYes: it is sent to the provider to make the audio
Voices102, plus the voices you clone1,108, with native speakers in 45 languages
LanguagesEnglish (US and UK), French, German, Spanish, Portuguese48, including Urdu, Hindi, Arabic and Italian
SpeedDepends on your phone or computer; the first use downloads the voice model onceDoesn’t depend on your device
Free5,000 characters a week with a free account1,000 characters once with a free account
Basic ($10 a month)Unlimited generations, up to 10,000 characters each100,000 characters a month
Voice cloningPrivate cloning in five languages, unlimited on BasicCloud cloning with a consent recording; up to 3 voices on Basic
DownloadsMP3 or WAVMP3 or WAV

Figures are KUKU Labs’ own plans, checked against what the app enforces. See pricing for every plan.

What “private” text to speech means

A private voice is a speech model that your browser downloads once and then runs itself. When you press generate, the work happens on your phone or computer: the script never reaches our servers. That is why it can be unlimited for a flat price, since nothing runs on a server we pay for each time.

The trade-offs are honest ones. The first use takes a moment while the model downloads, long scripts are faster on a newer device, and the private voices cover five languages. For scripts you would rather not share, such as unreleased videos, client work or anything personal, it is the simplest choice.

What “cloud” text to speech means

A cloud voice runs on a speech provider’s servers. Your text is sent there, the audio comes back, and your device only plays it. That gives you far more voices and languages, steady speed on any device, and voices recorded from native speakers, such as our Urdu, Hindi and Arabic voices.

Because every character costs us something at the provider, cloud voices come with a monthly allowance instead of being unlimited. We keep how many characters you used, never the text; the Privacy Policy explains what the provider receives.

Which should you choose?

  • Choose private for long English, French, German, Spanish or Portuguese narration you rewrite often: retakes are free and nothing is uploaded.
  • Choose cloud for Urdu, Hindi, Arabic and the other premium languages, or when you want a particular accent.
  • Choose cloud on an older phone or computer if private generation feels slow.
  • Mix them: a video can use a private voice for the narration and a premium voice for a second language.

Questions to ask any text to speech tool

  • Is my text sent to a server, and is it kept?
  • Is “unlimited” really unlimited, or is there a monthly character or credit allowance?
  • Which formats can I download, and can I use the audio in monetised videos?
  • Can I hear every voice before I pay?

You choose, every time.

The private tools and the cloud tools are separate in the Studio, and nothing is sent to the cloud provider unless you pick a premium voice and press generate.

Questions

Is private text to speech really private?

Yes, for the private voices: the speech is made in your browser and your text isn’t uploaded to our servers. The premium cloud voices are different: they send the text to our cloud speech provider.

Which sounds better?

It depends on the voice and the script, so listen for yourself: every voice has a free sample. The premium cloud voices offer more accents and native speakers in 48 languages.

Why is private unlimited but cloud isn’t?

Private voices run on your own device, so they cost us nothing per character. Cloud voices run on a provider we pay per character, so Basic includes 100,000 cloud characters a month and Pro 2,000,000.

Can I try the cloud voices free?

Yes. Every sample is free to hear, and a free account can generate 1,000 characters once with any premium library voice.

Does private text to speech work offline?

Not fully. The Studio needs an internet connection to open and, on the free plans, to count your characters, but the speech itself is made on your device once the voice model has downloaded.

Ready when you are.

Text to speech needs no account to try. A free account gives you 5,000 characters a week and unlimited voice enhancing.

Try it — no sign-up