30 voices, plus 2,000+ regional ones
Start with 30 studio voices, or pick a library voice with a regional accent, such as Mexican Spanish or Liverpool English.
Gemini 3.8 Flash TTS: Gemini text to speech with one or two voices
Give Manus your script and get natural speech back as a WAV file. Choose the voices, add a second speaker, and set the tone for each line.
0:00
0:23
Small Wins podcast intro, two voices
Why Gemini 3.8 Flash TTS
Gemini TTS with more voices and direction for every line
One speaker or a two-person dialogue
Give two speakers their own voices in one file, for podcasts, interviews, and practice calls.
A different tone on each line
Split the script into lines and give each one its own delivery, like calm, whispered, or excited, with sounds like a laugh or a sigh.
A closer look
What you can make with Gemini 3.8 Flash TTS
Every clip below was made with Gemini 3.8 Flash TTS in Manus. Each card shows the exact script, voices, and tone.
0:00
0:22
Lighthouse scene, Achernar voice
Audiobook scenes that change tone
One voice goes from calm to a nervous whisper to panic, because each line has its own direction.
Prompt
Use Gemini 3.8 Flash TTS with the Achernar voice to read this audiobook scene in three lines. Line 1, calm storytelling: "The lighthouse had been dark for eleven years. So when Nora saw the lamp flicker on, just after midnight, she told herself it was the storm." Line 2, whispered and nervous: "Then she heard the footsteps on the stairs above her... <short pause> slow, and getting closer." Line 3, panicked and speaking fast: "No -- no, no, the door is locked from the OUTSIDE! Who's up there?"
0:00
0:24
Support call, two library voices
Two-person calls with regional voices
A practice support call between two British English library voices, one from Winchester and one from Liverpool.
Prompt
Use Gemini 3.8 Flash TTS to voice this support call with two British English library voices: Hannah is en-gb-csagent-6 and Sam is en-gb-csagent-5. Hannah, professional and friendly: "Thanks for calling Harbour Line Rail, this is Hannah. How can I help today?" Sam, a bit stressed: "Hi, yeah, my train to Leeds was cancelled |oh no, I'm sorry| and I've got a meeting at ten. Is there anything earlier?" Hannah, professional and friendly: "Let me check... <short pause> Right, there's an eight fifteen from platform four. I can move your ticket onto it now, at no extra cost." Sam, relieved: "Oh, brilliant. You've saved my morning."
0:00
0:15
Coffee ad, Mexican Spanish
Ads in Mexican Spanish
A library voice with a Mexican Spanish accent reads a 15-second coffee ad, warm and confident.
Prompt
Use Gemini 3.8 Flash TTS with the Mexican Spanish library voice es-419-commercial-8 to read this coffee ad, warm and confident: "¿Mañanas difíciles? Prueba Café Alba: tostado en pequeños lotes y listo en tres minutos. Con notas de cacao y caramelo, cada taza sabe a domingo, aunque sea lunes. Café Alba. Despierta con calma."
0:00
0:23
Guided relaxation, Chinese
Calm narration in Chinese
A warm studio voice reads a guided relaxation script slowly, pausing where the script pauses.
Prompt
Use Gemini 3.8 Flash TTS with the Sulafat voice to read this guided relaxation script in Chinese, warm, calm, and slow: "欢迎来到今天的五分钟放松练习。先找一个舒服的姿势，轻轻闭上眼睛。吸气……慢慢地，数到四。呼气……让肩膀一点一点放松下来。很好，我们再来一次。"
Specs
Gemini 3.8 Flash TTS specs in Manus
Each request takes up to 6,000 characters of script and one or two speakers, and returns a WAV file.
How it works
How to use Gemini text to speech in Manus
01
Open Audio
Choose Audio in Manus. Every speech request in Manus uses Gemini 3.8 Flash TTS, so there is no model to pick.
02
Paste or describe your script
Paste the words to read, or ask Manus to write them. Say which voices to use and how each line should sound.
03
Listen and ask for changes
Play the WAV file in the chat, then ask for a different voice, pace, or tone. Long scripts are made in parts and joined into one file.
Choose the right model
Gemini 3.8 Flash TTS or another audio model
Manus has a model for speech, one for music, and one for sound effects. Here is when to pick each one.
Lyria 3.5
Pick it for music, from background tracks to songs with vocals, up to about three minutes long.
ElevenLabs
ElevenLabs Sound Effects v2
Pick it for sound effects, like rain, footsteps, or a door closing, from 0.5 to 30 seconds long.
More models in Manus
Compare every model in Manus, from images and video to music and sound.
Často kladené otázky
What is Gemini 3.8 Flash TTS?
Gemini 3.8 Flash TTS is a Google TTS model from the Gemini TTS family. In Manus it turns a script of up to 6,000 characters into natural speech with one or two voices and saves it as a WAV file.
How do I use Gemini 3.8 Flash TTS in Manus?
Open Audio in Manus, paste or describe your script, and say which voice and tone you want. Every speech request in Manus uses Gemini 3.8 Flash TTS, so you do not need to pick a model.
How does pricing work for Gemini 3.8 Flash TTS in Manus?
Availability and credit usage depend on your Manus plan. See pricing
How is Gemini 3.8 Flash TTS different from Gemini 3.1 Flash TTS?
Gemini 3.8 Flash TTS replaced Gemini 3.1 Flash TTS in Manus. Scripts now go in line by line, each line can have its own tone, and besides the 30 studio voices you can pick from more than 2,000 library voices.
Should I use Gemini 3.8 Flash TTS, Lyria 3.5, or ElevenLabs Sound Effects v2?
Use Gemini 3.8 Flash TTS for spoken words, from voice-overs to two-person dialogue. Use Lyria 3.5 for music and ElevenLabs Sound Effects v2 for sound effects.
How many voices and languages does Gemini 3.8 Flash TTS have?
Manus offers 30 studio voices that work in every supported language, plus more than 2,000 library voices with regional accents across 30 locales. Write the script in the language you want, such as Chinese, Japanese, or Spanish. Gemini 3.8 Flash TTS reads it as written and does not translate.
Can Gemini 3.8 Flash TTS make a conversation between two people?
Yes. One request can have two speakers, each with their own voice. For three or more speakers, Manus makes the script in parts of up to two speakers and joins them.