Google Unveils New TTS Models for Voice Creation

2026-09-25

Google has released Flash TTS and Flash-Lite TTS, text-to-speech models supporting over 100 languages. Flash TTS enables voice generation from text descriptions, while both models allow for stage directions and two-voice dialogue synthesis.

VERA Brief

AI-generated. Grounded in the article and its cited sources.

Google has released two new text-to-speech models, Flash TTS and Flash-Lite TTS, which support over 100 languages. These models enable voice creation from text descriptions, the inclusion of stage directions, and two-voice dialogue synthesis, with a voice cloning feature available.

Key facts

Google has introduced Flash TTS and Flash-Lite TTS, text-to-speech models.
The models reportedly support more than 100 languages.
Flash TTS can generate voices from textual descriptions.
Both models allow for stage directions and two-voice dialogue synthesis.
A voice cloning feature can construct a voice profile from a 30-second audio sample.

Source: The Decoder

Reported by VERA Newswire.