Text to Speech API: free key, plans from $5

Text to speech API with a free key and no card: any catalog voice, 31 languages, streaming from 300 ms. Plans from $5 a month for up to 216 hours of audio.

Voice texts from your own programs: a long text is split between the workers, and the file is ready in seconds.

A free key with no card: 3 minutes of audio a day and any catalog voice. Plans from $5 a month for up to 216 hours of audio; first sound from 300 ms, 31 languages.

Start for free

3 min a day

Free 3 min a day

Starter 3 min every 10 min

Author 8 min every 10 min

Pro 20 min every 10 min

Studio 65 min every 10 min

Plans

A dearer plan starts at once, and the days left of the current one are converted into days of the new one. A cheaper plan starts when the current one ends. Every catalog voice works without buying it on any plan, even the free one: plans differ only in the number of seconds.

API keys

The key is free and needs no card: sign in with Google or GitHub and create it here. Keys and plans belong to your account.

API reference

Base address: https://superclonica.com/api/v1. Every request goes over HTTPS with an Authorization: Bearer header and your key. The key is created above and shown once: keep it on your server, not in a web page’s code.

Methods

Voice a text

The request body is JSON. An unknown field is refused with a 422 error.

In Python, retrying on 429 and 503, with an idempotency key so a retry never voices the text twice:

In JavaScript (Node.js):

Streaming

With "stream": true the text is split into pieces, they are voiced on several workers at once, and the server sends the sound as soon as the first piece is ready: the first sound from 300 ms whatever the length of the text. Ogg always streams; mp3 and wav stream when the text is longer than about half a minute of speech. If a stream stops halfway, treat the request as failed and send the text again.

Voices and what is left of the plan

kind: stock — the standard voices, own — yours, bought — bought ones, catalog — catalog voices (available on any plan, even the free one).

Headers

Answers and errors

An error comes as JSON with a detail field. A failed text does not use the plan.

Limits

A plan gives seconds of audio in every 10-minute window; the free plan 180 seconds a day.

One request takes up to 20,000 characters, no longer than the plan’s window and no more than 30 minutes of audio (wav: 15 minutes).

Up to 20 unfinished tasks per account at once.

A text not voiced within 85 seconds is cancelled with a 503 answer and does not use the plan.

Frequently asked questions

Is the API key really free?

Yes. Sign in with Google or GitHub and create a key on this page; no card is needed. The free plan gives 3 minutes of audio a day and never runs out.

Which voices can the API use?

Every catalog voice, on any plan, the free one too, with no purchase. Plus the 10 standard Supertonic 3 voices and the voices you train on the site from your own recording.

How do the plans differ?

Only in the seconds of audio in every 10-minute window: Starter for $5 gives 3 minutes, Author for $10 gives 8, Pro for $20 gives 20 and Studio for $50 gives 65. Voices, languages and speed are the same on all of them.

How do I pay for a plan?

By card via Getly right at checkout, or from the site balance, which is topped up in cryptocurrency via NOWPayments. A plan runs for 30 days; paid from the balance, it can renew itself.

How fast does the audio arrive?

The answer streams: the first sound comes from 300 ms in our measurements. A long text is split between workers, so 15 minutes of speech is ready in about 25 seconds.

What happens when the limit runs out?

The API answers 429 with a Retry-After header that says in how many seconds the request will fit. Nothing is charged, and a failed text does not use up the plan.

Loading SuperClonica…