Multilingual text to speech

Definition

Multilingual text-to-speech generates spoken audio in different languages from one API. SpeechifyAI’s Simba 3.5 speaks 48 languages under one model id, simba-3.5, on every plan. Every voice in the catalog speaks all 48, and so do your cloned voices on a paid plan. It lets a product reach a global audience without integrating a separate voice service per language: you translate the text, pick a voice, and synthesize.

The 48 languages

Arabic, Belarusian, Bengali, Bosnian, Burmese, Catalan, Chinese, Croatian, Czech, Danish, Dutch, English, Finnish, French, Georgian, German, Greek, Gujarati, Hebrew, Hindi, Hungarian, Indonesian, Italian, Japanese, Kazakh, Korean, Latin, Malay, Marathi, Norwegian, Persian, Polish, Portuguese, Romanian, Russian, Serbian, Slovak, Slovenian, Spanish, Swedish, Tagalog, Tamil, Telugu, Thai, Turkish, Ukrainian, Urdu and Vietnamese. Any regional variant of each is in scope, so pt-BR and pt-PT are both Portuguese. The older Simba 3.0 covers 6 languages and Simba 3.2 is English only; both remain available.

Localization as synthesis

Because everything is text-driven, localizing content is a synthesis job. An audiobook, a course library, a set of IVR prompts, or in-app notifications can each be produced in another language by re-synthesizing the translated text. Nothing needs re-recording.

One integration

The same endpoint and the same model id serve all 48 languages, so you integrate once and set language per request. That keeps a multi-market product simple instead of stitching together per-language vendors.

Natural voices

For the best results, use a voice native to the language. A voice keeps the accent of the recording it was made from. Cloned voices work on Simba 3.5 on a paid plan. See every Simba 3.5 language and check the model guide before integrating.