How Speakvora Text-to-Speech Billing Works

Speakvora charges per character for text-to-speech: $0.70 per 1M for catalog voices, $6 per 1M for custom voices. Library hits are billed every time.

Short answer. Speakvora bills text-to-speech by character count. Catalog voices cost $0.70 per 1 million characters, custom voices cost $6 per 1 million characters, and library hits are billed by character every time. Each account receives 100,000 free characters.

Billing Unit: Characters

Speakvora charges for text-to-speech by the number of characters in your request text. Each time you call POST /v1/speak, the character count of the text parameter is billed, regardless of whether the audio is generated fresh or retrieved from the library.

Catalog Voice Pricing

Catalog voices (Kokoro voices like af_heart, ef_dora, em_alex, ff_siwis, pf_dora, pm_alex, hf_alpha, hm_omega, jf_alpha, jm_kumo, zf_001, and zm_009) cost $0.70 per 1 million characters. This rate applies whether the text is already in the library or generated live for the first time.

Custom Voice Pricing

Custom voices that you record, upload, or design in the dashboard cost $6 per 1 million characters. If you add a delivery style (happy, excited, sorry, calm, serious—English only), the rate increases to $17 per 1 million characters, but only if the style is successfully applied; if the backup engine responds without the style, the plain $6 rate applies instead.

Library Hits Are Billed Every Time

When you request text that already exists in the library, it returns instantly but is still billed at the full character rate. If you speak the same phrase 10 times, you are billed 10 times. There is no special rate for repeated requests.

Preparing Text Without Billing

Use POST /v1/prepare to queue up to 200 texts for generation in the background without being billed. The endpoint returns how many texts are ready (already in the library) and how many are queued for generation. Queued texts become available in a few seconds, and subsequent /v1/speak requests for those texts are billed normally. This is useful for warming up the cache before you need the audio.

Free Allowance Per Account

Every account receives 100,000 free characters. After you exceed this allowance, all additional characters are billed at the rates above. The free allowance applies to all characters regardless of whether they come from the library or are generated live.

Other Billable Services

Speech-to-text is billed separately at $0.0015 per minute of audio (60 free minutes per account). Voice agents are billed at $10 per 1,000 turns (200 free turns per account). Offline packs are one-time purchases.

Frequently asked questions

Do I pay for library phrases every time I use them?

Yes. Library hits are billed by character count each time you request them, even if the audio already exists. There is no special rate for repeats.

What does POST /v1/prepare do and does it cost anything?

POST /v1/prepare queues up to 200 texts for background generation without billing. It returns how many are ready and how many are queued. You are billed only when you later call /v1/speak for those texts.

How much does a custom voice cost compared to a catalog voice?

Custom voices cost $6 per 1 million characters, compared to $0.70 per 1 million for catalog voices. Styled custom voices (English only) cost $17 per 1 million if the style is applied.

What is the free allowance and how does it work?

Every account receives 100,000 free characters. This allowance applies to all characters you generate, whether from the library or live. After you use it, all additional characters are billed at the standard rates.

Are speech-to-text and voice agents billed the same way?

No. Speech-to-text is billed at $0.0015 per minute of audio (60 free minutes per account). Voice agents are billed at $10 per 1,000 turns (200 free turns per account). Each service has its own free allowance.