## Documentation Index

Fetch the complete documentation index at: [/llms.txt](https://docs.cognigy.com/llms.txt)

Use this file to discover all available pages before exploring further.

The `say` verb converts text to speech and plays it to the caller. It supports multiple TTS vendors, languages, voices, and advanced options such as looping, early media, and fallback synthesizers.

```
{
  "verb": "say",
  "text": [\
    "Hi there!",
    "Welcome to our service."
  ],
  "synthesizer": {
    "vendor": "google",
    "label": "primary-tts",
    "language": "en-US",
    "voice": "en-US-Wavenet-F",
    "engine": "neural",
    "gender": "FEMALE",
    "fallbackVendor": "aws",
    "fallbackLabel": "fallback-tts",
    "fallbackLanguage": "en-GB",
    "fallbackVoice": "Amy",
    "options": {
      "speakingRate": 1.0,
      "pitch": 0.0,
      "volumeGainDb": 0.0
    }
  },
  "loop": 2,
  "earlyMedia": true,
  "disableTtsCache": false
}
```

## Configuration

The following table lists the available parameters:

| Parameter                        | Type                   | Description                                                                                                                                                                             | Required |
| -------------------------------  | ---------------------- | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | -------- |
| text                            | string \| array       | The text to speak. The text may be either plain text or use SSML tags for advanced speech effects such as pauses, emphasis, or embedded audio. Can be a string or an array of strings. For example, `<speak>Hello, <break time='1s'/> welcome!</speak>`. | Yes      |
| synthesizer                     | string \| object      | Defines the text-to-speech options. Can be an object with detailed TTS settings or a string referencing a pre-configured synthesizer.                                                    | No       |
| synthesizer.vendor              | string                 | The TTS service provider to use, for example, `google` or `aws`. This determines the available voices, languages, and engine options. See [supported speech vendors](https://docs.cognigy.com/voice-gateway/references/tts-and-stt-vendors) or [add a custom speech API](https://docs.cognigy.com/voice-gateway/webapp/speech-services#add-custom-speech-vendors). | No       |
| synthesizer.label               | string                 | A custom label for this synthesizer instance. Useful for logging or identifying multiple TTS configurations in a single flow. Can be `null` or a string.                              | No       |
| synthesizer.language            | string                 | The language code for the speech output, for example `en-US` or `de-DE`. Required if a vendor is defined.                                                                            | No       |
| synthesizer.voice               | string \| object      | The specific voice to use. Can be a string representing the vendor-specific voice name or an object with advanced properties. Defaults to the Application-level TTS voice if not provided. For example, `en-US-Wavenet-F` for Google TTS. | No       |
| synthesizer.engine              | string                 | The TTS engine type. Options are: <br>- **standard** – the default engine<br>- **neural** – a high-quality natural voice<br>- **generative** – an experimental AI voice<br>- **long-form** – optimized for long text | No       |
| synthesizer.gender              | string                 | The desired voice gender: `MALE`, `FEMALE`, or `NEUTRAL`. Used for vendors that support gender selection.                                                                              | No       |
| synthesizer.fallbackVendor      | string                 | An alternative TTS vendor to use if the primary vendor fails or returns an error. For example, `aws`.                                                                                  | No       |
| synthesizer.fallbackLabel       | string                 | A label for the fallback TTS instance. Helps distinguish the primary and fallback synthesizers in logs.                                                                                   | No       |
| synthesizer.fallbackLanguage    | string                 | The language code for the fallback synthesizer, for example `en-US` or `de-DE`. Defaults to the primary language if not provided.                                                       | No       |
| synthesizer.fallbackVoice       | string \| object      | The voice for the fallback synthesizer. Can be a string or object. For example, `Amy` for AWS.                                                                                        | No       |
| synthesizer.options             | object                 | A vendor-specific TTS options object, such as `speakingRate` from 0.25 to 4.0, `pitch` from -20 to 20, or `volumeGainDb` from -96 to 16. These options control the speech speed, pitch, and volume. | No       |
| loop                            | number                 | The number of times to repeat the utterance. Set 0 for infinite repetition. The default value is 1. Useful for hold music or repeated prompts.                                          | No       |
| earlyMedia                      | boolean                | If true and the call isn’t yet answered, plays the audio without officially answering the call. Useful for IVR previews or ringback messages. The default value is false.                | No       |
| disableTtsCache                | boolean                | If true, disables caching of the synthesized audio for this utterance. The default value is false.                                                                                     | No       |