> For clean Markdown of any page, append .md to the page URL.
> For a complete documentation index, see https://docs.jambonz.org/llms.txt.
> For AI client integration (Claude Code, Cursor, etc.), connect to the MCP server at https://docs.jambonz.org/_mcp/server.

# Google

Google tuning is passed in `synthesizer.options`.

```json
{
  "verb": "say",
  "text": "Thanks for calling, how can I help you today?",
  "synthesizer": {
    "vendor": "google",
    "language": "en-US",
    "voice": "your-gemini-voice-name",
    "options": {
      "model": "gemini-2.5-flash-preview-tts",
      "prompt": "Say this in a warm, friendly tone."
    }
  }
}
```

## Options

### Gemini TTS

**`model`** `string`

The model to use. When set, this enables Gemini TTS. Example: `gemini-2.5-flash-preview-tts`.

---

**`apiMode`** `string`

Which Google TTS API to use: `tts` (standard Cloud TTS voices), `live` (HD voices in streaming mode) or `gemini` (Gemini TTS). When omitted it is chosen for you: `gemini` if a model is set on the verb or credential, `live` for an HD voice, otherwise `tts`.

---

**`prompt`** `string`

Guidance on tone, emotion or speaking style for the generated speech. Gemini TTS only.

---

> **Note**
>
> These options apply to the `say` verb. Google TTS used by an [agent](/verbs/verbs/agent) over the streaming path uses only the voice and language.

## Streaming

Google supports [TTS streaming](/guides/features/tts-streaming).