Text to Speech

Convert text into lifelike audio using Google's Gemini TTS model. Supports multilingual synthesis, 30 distinct voice personas, and optional style prompting to control tone, pace, and delivery.

Available Voices

30 voice personas across male and female options with varying pitch and character. Here are some popular picks:

Voice	Gender	Pitch	Character
Zephyr	Female	Higher	Bright
Puck	Male	Middle	Upbeat
Charon	Male	Lower	Informative
Kore	Female	Middle	Firm
Aoede	Female	Middle	Breezy
Fenrir	Male	Lower middle	Excitable
Leda	Female	Higher	Youthful
Orus	Male	Lower middle	Firm
Achernar	Female	Higher	Soft
Algieba	Male	Lower	Smooth
Sulafat	Female	Middle	Warm
Achird	Male	Lower middle	Friendly

Use GET /studio/speech/voices in the API Playground below to browse all 30 voices.

Style Prompting

Use the style_prompt parameter to control how the voice delivers the text. Write natural-language instructions describing the desired tone.

Examples:

• "Speak in a calm, authoritative news anchor tone"

• "Read slowly and dramatically, like an audiobook narrator"

• "Sound cheerful and enthusiastic, like a podcast host"

• "Whisper softly, as if telling a secret"

Multilingual Support

Pass a language_code (BCP-47 format) to synthesize speech in a specific language. Use GET /studio/speech/languages to retrieve all supported languages.

Common codes: en-US, hi-IN, fr-FR, es-ES, de-DE, ja-JP, ko-KR, zh-CN, ar-SA, pt-BR.