Voice model by Google DeepMind
Gemini 3.8 Flash TTS
Google's directable text-to-speech model. Describe the voice you want in plain words, from character and accent to tone and pacing, and Craze narrates your video with it.
- Makes
- Text to audio
- Starts at
- $0.032 per 1,000 characters
- Plans
- Every plan, including Free
Gemini 3.8 Flash TTS is Google's text-to-speech model for narration you direct in plain words. You choose a voice, write how it should sound, and the model performs the script to match: the character, the accent, the tone and the pace.
Why creators choose Gemini 3.8 Flash TTS
Most text-to-speech models give you a list of voices and little else. Gemini 3.8 Flash TTS treats the voice as a performance you can direct, so a single stock voice can become a hushed storyteller, a brisk sports announcer or a tired detective.
Design a voice by describing it
Write a voice direction next to your script: who is speaking, how old they sound, where they are from, how they feel and how fast they talk. The direction steers the read and is never spoken, so the script stays exactly as you wrote it.
Thirty voices to start from
The direction works on top of 30 prebuilt voices, from firm and informative to breezy and bright. Pick the one closest to what you hear in your head, then shape it with the direction.
Vocal moments in the script
Mark small vocal events inline, such as a laugh or a sigh, where they belong in the line.
How to use Gemini 3.8 Flash TTS in Craze
- Write or paste your script.
- Name Gemini 3.8 Flash TTS in chat along with how the voice should sound, or pick it in the Models panel, choose a voice and fill in the voice direction.
- Craze generates the narration and adds it to your project.
- Trim it in the timeline, or change the direction and generate another take.
Gemini 3.8 Flash TTS pricing
Pricing starts at $0.032 per 1,000 characters. The cost follows the length of the script, not the voice or the direction. Use the calculator on this page to estimate the cost before you generate.
This model is available on every Craze plan, including Free.
Gemini 3.8 Flash TTS vs other models
Gemini 3.8 Flash Lite TTS takes the same voices and directions at a lower price, and suits drafts and short lines. Gemini 3.1 Flash TTS is the earlier Gemini voice model. ElevenLabs v4 is Craze's default narrator, and it can also save a designed voice to reuse across projects.
Limitations
On Craze, a script can run to 8,000 characters per generation; split longer narration into sections and join them in the timeline. The model returns voice only, so music and sound effects come from other tracks or generations. A voice direction applies to one generation; to reuse a sound, keep the same voice and direction.
Gemini 3.8 Flash TTS specs on Craze
| Task | Inputs | Resolution | Aspect ratios | Sound |
|---|---|---|---|---|
| Text to speech | Text prompt | Follows the source | Follows the input | None |
Frequently asked questions
Gemini 3.8 Flash TTS is a text-to-speech model from Google DeepMind. You pick one of 30 voices and describe how it should sound in plain words, such as its character, accent, tone and pacing, and the model performs your script that way.
You pay $0.032 per 1,000 characters. The cost calculator on this page estimates the total before you generate.
every Craze plan, including Free.
Yes. Write a voice direction alongside your script, for example a warm, older British narrator reading slowly, like a bedtime story. The direction shapes the performance and is never read aloud. Change it and generate again to try another take.
Both take the same voices and plain-word directions. Flash Lite TTS is the lighter, cheaper of the two, suited to quick lines and drafts; Flash TTS is the one to reach for on finished narration.
Name Gemini 3.8 Flash TTS in chat with your script and how the voice should sound, or pick it in the Models panel, choose a voice, add a voice direction and generate. Craze adds the narration to your project, ready to lay over video.
