Prompt-based voice design
Gemini 3.8 Flash TTS can create bespoke voices by prompting for a role, accent and other vocal characteristics across more than 100 languages and dialects.
Gemini 3.8 Text-to-Speech
Gemini 3.8 Text-to-Speech turns scripts into customizable spoken audio with expressive voice design, line-by-line direction and dialogue generation.
Overview
Gemini 3.8 Text-to-Speech is a pair of expressive audio-generation models from Google: Gemini 3.8 Flash TTS and Gemini 3.8 Flash-Lite TTS. They turn text and scripts into directed speech, moving beyond fixed voice presets with prompt-based voice design, detailed performance instructions and support for multi-speaker scenes.
Flash TTS is intended for creative voice and character design, while Flash-Lite TTS is positioned for high-volume dubbing, audio creation and expressive voice agents. The source lists Google AI Studio, the Gemini API, Gemini Enterprise, Gemini Notebook and Google Vids as product contexts for these models, although it does not provide detailed access, pricing or quota information.
Key features
Large voice library
The source describes a library of more than 2,000 production-ready voices, including regional varieties such as Mexican Spanish, Quebec French and Scots English.
Line-by-line performance control
Users can direct each line with natural-language stage directions for pacing, emotion, dialect shifts, delivery style and conversational reactions.
Two-speaker scene staging
Scripts can stage two speakers in a single scene while preserving distinct voices and natural conversational turn-taking.
Long-form generation
The models support extended audio generation for podcasts and audiobooks, with natural pacing and character timbre intended to remain consistent over hours of audio.
Voice replication and provenance tools
Voice replication can use a 30-second sample of a voice the user owns or has rights to use. The source also cites consent verification, SynthID watermarking and C2PA credentials for this workflow.
Use cases
Character voices and interactive media
Design a recurring character or narrator with a prompted vocal identity, then direct individual lines for dramatic scenes, games or immersive storytelling.
Audiobooks and podcasts
Produce long-form spoken content with controlled pacing and consistent character timbre for audiobooks and podcasts.
Dubbing and content localization
Create localized or high-volume spoken content using the Flash-Lite model’s positioning for dubbing and expressive audio production.
Voice agents
Build voice agents that use tone, pacing, backchanneling and conversational reactions to make spoken interactions more expressive.
Two-person dialogue scenes
Write a multi-turn script with two distinct speakers and control their handoffs and reactions for narrative or podcast-style scenes.
Pros and Cons
Pros
- Supports custom voice creation from natural-language prompts, including role, accent and vocal characteristics.
- Provides detailed script control for pacing, emotion, nonverbal cues and conversational backchanneling.
- Offers two-speaker scene staging and long-form generation for extended narrative audio.
- Includes a broad voice library and language coverage spanning more than 100 languages and dialects.
- Describes consent verification, SynthID watermarking and C2PA credentials for voice-replication workflows.
Cons
- The source does not document pricing, quotas, output formats, latency or detailed availability requirements.
- Voice remixing—fine-tuning an existing library voice’s timbre, pitch, pace or accent—is listed as coming soon rather than generally available.
- Voice replication should be used only with a sample the user owns or has permission to use; the source does not provide further operational or legal guidance.
FAQ
What is Gemini 3.8 Text-to-Speech?
The product includes two models: Gemini 3.8 Flash TTS for deeper creative direction and custom character voices, and Gemini 3.8 Flash-Lite TTS for high-volume, cost-efficient generation. Both support line-by-line performance direction.
Where can the models be used?
The source lists Google AI Studio, the Gemini API, Gemini Enterprise, Gemini Notebook and Google Vids as places where the models can be used. Specific availability, account requirements and regional restrictions are not provided in the source.
Can users create or replicate a voice?
Yes. Flash TTS can design voices from natural-language prompts, and the source describes voice replication from a 30-second sample of a voice that the user owns or has permission to use. Consent verification, SynthID watermarking and C2PA credentials are described as part of the replication workflow.
How can generated speech be directed?
Users can direct delivery line by line with stage directions and script cues, including pacing, emotion, dialect changes, vocal bursts and backchanneling. The models also support native two-speaker scene staging and long-form generation with voice consistency across extended audio.
Quick Facts
- Product type
- Text-to-speech and expressive audio generation models
- Models
- Gemini 3.8 Flash TTS and Gemini 3.8 Flash-Lite TTS
- Primary users
- Creators, developers and enterprises
- Language coverage
- More than 100 languages and dialects
- Listed product contexts
- Google AI Studio, Gemini API, Gemini Enterprise, Gemini Notebook and Google Vids
- Source domain
- blog.google
Gemini 3.8 Text-to-Speech Alternatives
Fylom
Fylom is an AI podcast generator that turns topics or questions into researched, narrated episodes with voice selection and recurring listening.
Gemini 3.1 Flash TTS
Gemini 3.1 Flash TTS is Google’s preview text-to-speech model for expressive AI speech with fine-grained style and delivery control across Gemini API, Google AI Studio, Vertex AI, and Google Vids.
蓝藻AI
蓝藻AI is an online AI voice generation and dubbing platform that turns text into speech and supports self-service voice cloning for short videos and audiobooks.
Ondoku
Ondoku is a browser-based text-to-speech tool that turns text into downloadable .mp3 audio, with free and paid plans, multilingual reading, image reading, and commercial use options.
Typecast
Typecast is an online AI voice generator that turns text into life-like speech with emotional delivery and hyper-realistic voices.
Noiz AI
Noiz AI is an AI text-to-speech, voice cloning, and voice design tool for lifelike speech from text, with emotion control in one workflow.