Available now

Gemini 3.8 Flash TTS: text to speech

By Google

Google's Gemini 3.8 Flash TTS reads a script in one of its prebuilt voices, or voices a two-person dialogue, with directions for tone kept apart from the words.

Model ID
Maker
Google
Added
September 29, 2026

Input

Fields marked * are required. Review the price before generating.

Verbatim text for single-speaker speech. For dialogue, omit prompt and provide speakers and turns instead. The complete input is limited to 8,192 tokens.

Example prompt. Edit it or write your own.

Additional settings

Delivery style, separate from the spoken transcript. Applies to all turns unless overridden.

Prebuilt voice for single-speaker speech.

Exactly two distinct speaker aliases and their prebuilt voices for dialogue.

Ordered dialogue turns, each identifying a configured speaker.

Price$0.08 / run

You sign up first. Your settings are kept.

About this model

Overview

Gemini 3.8 Flash TTS turns text into speech. For one speaker, write the exact words in the prompt, pick a voice such as Kore, Puck or Charon, and describe the delivery in the style field. Sounds like <laugh> or <sigh> can sit inside the text.

For a dialogue, leave the prompt empty, give two speakers a name and a voice each, then list the turns in order. Each turn can carry its own style direction on top of the shared one.

What you can build

  • Product walkthrough narration with a set tone
  • Two-host podcast intros and short interviews
  • Character exchanges for games and animations
  • Audio lessons with a teacher and a student voice

Prompting tips

  • Keep directions out of the script: "calm, slow, friendly" goes in the style field, not the prompt.
  • Give a single turn its own style when only one line should change, such as a whispered aside.
  • Pick two voices that sound clearly different so listeners can tell the speakers apart.

What to check

Listen for omitted words, pronunciation of names and numbers, unexpected pauses and any directions spoken aloud. In dialogue, check that every line belongs to the intended speaker. Align the finished audio with the video in your editor; a written pacing instruction does not guarantee an exact duration.

Inputs

  • Prompt (optional): Verbatim text for single-speaker speech. For dialogue, omit prompt and provide speakers and turns instead. The complete input is limited to 8,192 tokens.
  • Style Instructions (optional): Delivery style, separate from the spoken transcript. Applies to all turns unless overridden.
  • Voice (optional): Prebuilt voice for single-speaker speech. Options: Achernar, Achird, Algenib, Algieba, Alnilam, Aoede, Autonoe, Callirrhoe, Charon, Despina, Enceladus, Erinome, Fenrir, Gacrux, Iapetus, Kore, Laomedeia, Leda, Orus, Pulcherrima, Puck, Rasalgethi, Sadachbia, Sadaltager, Schedar, Sulafat, Umbriel, Vindemiatrix, Zephyr, Zubenelgenubi.
  • Speakers (optional): Exactly two distinct speaker aliases and their prebuilt voices for dialogue.
  • Turns (optional): Ordered dialogue turns, each identifying a configured speaker.

Output and limits

  • Sung or spoken words can come out unclear or differ from your text.
  • Exact length, tempo and key are not guaranteed; listen to the whole result before you use it.
  • Prompts that name real artists or protected songs can be refused.
  • Some prompts or outputs can be refused by safety checks.

Price guide

  • up to · 1,000 ASCII characters: $0.08 / run