Ten-language voiceover: cost of a 600-character script in TTS

Voicing a 600-character script in ten languages costs $0.285 on Sume, assuming each translated script is also about 600 characters. The pricing page, read 2026-10-03, lists text to speech at $0.0475 per 1,000 characters, so one version is $0.0285. The catch is that length changes by language, and the price follows characters, not seconds.

What does a multi-language set cost?

Spaces and punctuation count toward the 1,000-character unit, according to the pricing note, so a script padded with line breaks costs more than its words suggest.

Why will your real total differ?

• Translations rarely keep the same character count. A script that runs longer in the target language costs more, and a language that packs more meaning into each character costs less.
• Translation itself is not in this total, and neither are retakes.
• Characters are not seconds. Two languages with the same character count can speak for different lengths, which matters when the voiceover must fit a fixed clip.

How do I keep the versions aligned?

Voice each language as its own job with its own idempotency key, such as promo-fr, and keep the job ids in a table. If a version runs long for a fixed-length clip, retake that language alone with a tightened script and a new key; the same key with a changed script returns 409 idempotency_conflict. When one video needs several languages in sequence, join the takes with Timeline audio concat, which accepts up to 20 parts.

What I could not verify

I did not confirm which languages and voices the TTS route lists today, and I did not test voice quality by language. Read the live model and voice catalog before promising a language to a client.