Text to speech API pricing: per-character rates explained

Text to speech API pricing is a rate per amount of text you convert, so your bill is the characters you send times the rate. Sume's TTS API charges $0.0475 per 1,000 characters, plus a 5.5% agent fee by default; spaces and punctuation count, and one request takes up to 20,000 characters.

The rate is read from the code behind API pricing. The request limits come from the TTS schema in the Sume API reference, the OpenAPI document behind the API reference docs, and billing behavior from Usage and Generation admission, read on 2026-09-29.

What counts toward the character count?

Every character of the transcript you send counts, including spaces and punctuation. When you compare TTS prices across vendors, check the unit first: a rate per character, per word, or per minute of audio gives different totals for the same script.

• One transcript holds 1 to 20,000 characters. Longer text is split into several requests, each billed for its own characters.
• Synthesized audio longer than 1,200 seconds fails with tts_duration_exceeded, and no credit is captured.
• The completed result carries character_count in the current code, so you can log what each job counted.
• The /pricing/api rate card has one Text to speech row. It has no separate rows per voice.

How much does text to speech cost at my volume?

Multiply your characters by the rate. The character counts below are examples; count your own scripts. One maximum request of 20,000 characters comes to $0.95 before the fee.

How long is the audio I'm paying for?

Sume bills characters, not seconds, and publishes no characters-per-second rate. generation_config.speed, a multiplier from 0.6 to 1.5, changes the pace without changing the characters you pay for. To know the length, send timestamps: { "words": true }; the current code then adds duration_seconds to the result. The text to speech time calculator post covers estimating a script's length before you pay for it.

How do I pay, and when am I charged?

Sume has subscription plans (Free $0, Pro $40 / month, Startup $120 / month, Scale $400 / month; see Pricing) and bills API calls at each model's published USD rate from one balance. Teams fund it with on-demand purchases from $10 to $1000 in the dashboard.

• At submit, Sume reserves the estimated amount. A successful job captures it; a failed job releases or refunds the reservation.
• GET /v1/usage?job_id=… sums what one job cost; its debited_usd field is what the wallet deducted.
• GET /v1/balance returns the USD balance left.
• A submit that can't reserve the estimate fails with 402 insufficient_credits before any work starts.

Does picking a TTS model change the price?

TTS 1.0 has no engine picker: model and model_id are rejected with 400. The TTS Router lets you pin a catalog model; read its price from GET /v1/tts-router/models before you switch, rather than assuming the TTS 1.0 rate. For the request itself, see Text to speech API.