One or two directed speakers
Render narration or a two-person scene in one output. Give each speaker a voice, profile, accent, style, and pace, then assign every turn.
Create AI video drafts directly in Img2Vid. Choose a compatible model, review price and plan eligibility, then generate in the same workspace.
Checking price and access…
Gemini 3.1 Flash TTS
Gemini 3.1 Flash TTS is Google's preview model for controllable narration and dialogue. On Img2Vid, define one or two speakers, choose from 30 voices, direct style, pace, accent, and character, then write an ordered performance with audio tags. The result is structured speech generation rather than a generic prompt box.
Hear Gemini 3.1 Flash TTS across six production scenarios
Every example contains audio generated by Gemini 3.1 Flash TTS through the configured KIE task API. Img2Vid packages the original speech inside a 16:9 waveform video for consistent comparison across narration, bilingual delivery, distinct speakers, character acting, and product guidance.
Direction
Zephyr voice, American accent, Promo/Hype style, and natural pace. A premium launch read that builds from composed confidence to an energetic invitation.
Core Capabilities
What Gemini 3.1 Flash TTS can control
Gemini 3.1 Flash TTS combines exact recitation with natural-language direction. Google documents 30 voices, more than 70 languages, up to two speakers, and audio tags. Img2Vid presents those capabilities as explicit controls.
Thirty preset voices
Gemini 3.1 Flash TTS exposes Google's 30 preset voices, including Zephyr, Puck, Kore, Fenrir, Aoede, and Charon. Preview them because names alone cannot describe a performance.
More than 70 languages
Generate speech across more than 70 languages with automatic language detection. Native reviewers should still check names, abbreviations, and code switching.
Inline audio tags
Use directions such as [whispers], [laughs], or [gasp] inside dialogue. Gemini 3.1 Flash TTS can change one moment without changing the global profile.
Scene and global tone direction
Describe setting, relationships, emotional arc, and tone separately from the exact words. Scene context complements per-speaker controls.
SynthID provenance
Google states that Gemini 3.1 Flash TTS audio includes SynthID. Creators still need consent, rights review, disclosure, and editorial checks.
Production Workflow
A deliberate Gemini 3.1 Flash TTS workflow
Strong speech starts with casting and direction. Img2Vid keeps those decisions visible so each revision can target one variable.
- Step 01
Cast the voices first
Choose one or two voices, then define role, energy, age impression, and delivery without imitating a real person.
- Step 02
Set performance boundaries
Select accent, style, pace, and temperature. Keep scene direction compatible with the dialogue instead of giving contradictory instructions.
- Step 03
Build the ordered dialogue
Assign each turn to a speaker and write the exact text. Img2Vid allows eight turns of 250 characters as a product guardrail, not a Google limit.
- Step 04
Listen, compare, and revise
Review pronunciation, speaker separation, timing, and emotion. Change one Gemini 3.1 Flash TTS control at a time to isolate improvements.
Practical Uses
Where Gemini 3.1 Flash TTS fits best
Gemini 3.1 Flash TTS suits controlled recitation rather than open microphone conversation. These uses benefit from a known script and explicit roles.
Podcast segments
Create scripted intros, recaps, explainers, or two-host prototypes. Separate profiles keep the conversation intentional.
Dialogue · Editorial
Direction
Two studio hosts, distinct energy, natural timing, no announcer exaggeration.
Audiobooks and serial fiction
Direct a narrator or short scene with atmosphere, pacing, and tags. Gemini 3.1 Flash TTS helps test prose before full production.
Narration · Story
Direction
Restrained narrator, intimate distance, slower pace, controlled suspense.
Product education
Turn onboarding and help scripts into clear guidance. Keep a stable global tone while tags emphasize key actions.
Onboarding · Training
Direction
Empathetic guide, concise phrasing, pause between each action.
Character prototypes
Test game dialogue and pitch scenes with contrasting speakers. Final casting still requires creative and rights review.
Games · Previsualization
Direction
Weary guardian and urgent traveler, clear tension, final whispered reveal.
Multilingual localization
Create multilingual review drafts or code-switching lessons. Native speakers should approve pronunciation and cultural tone.
Localization · Learning
Direction
Patient teacher, clear Mandarin phrase, natural English response.
Branded status narration
Generate scripted weather, accessibility, or in-app updates. Verify dynamic facts before they reach Gemini 3.1 Flash TTS.
Product · Accessibility
Direction
Calm neutral delivery, factual language, no dramatic emphasis.
Model Positioning
How Gemini 3.1 Flash TTS differs from other speech workflows
Compare workflow fit, not unsupported absolute rankings. Gemini 3.1 Flash TTS emphasizes Google-native control and multilingual scripts; other platforms may prioritize cloning or streaming.
| Decision | Gemini 3.1 Flash TTS | ElevenLabs | PlayHT |
|---|---|---|---|
| Primary fit | Directed scripted narration and two-speaker scenes | Broad voice platform and voice design workflows | Streaming and conversational speech workflows |
| Current Img2Vid controls | 30 voices, profiles, accent, style, pace, tags | Not configured on this page | Not configured on this page |
| Speaker structure | One or two speakers in a single ordered request | Varies by selected product and model | Varies by selected model and endpoint |
| Voice cloning here | Not exposed by this integration | Available in its broader platform | Available in its broader platform |
FAQ
Gemini 3.1 Flash TTS questions
Practical answers about the model and the exact Gemini 3.1 Flash TTS configuration currently available on Img2Vid.
Create on Img2Vid
Direct your first Gemini 3.1 Flash TTS performance
Define each voice, write the dialogue, and use Gemini 3.1 Flash TTS to turn a script into expressive audio.