GenerateSpeechOptions

Options for generateSpeech.

Extends

• SpeechCommonOptions

Properties

authContextRef?

optional authContextRef?: string

Secure-storage slot for provider credentials. Never a raw secret.

Inherited from

SpeechCommonOptions.authContextRef

---

destinationRef?

optional destinationRef?: string

Opaque application destination, resolved by the host. Never a path or URL.

---

execution?

optional execution?: TextExecutionOptions

Inherited from

SpeechCommonOptions.execution

---

locale

locale: string

BCP 47 tag, e.g. "de-AT".

Inherited from

SpeechCommonOptions.locale

---

localeMatch?

optional localeMatch?: LocaleMatch

"language" (default) requires the language and prefers the region, so de-AT may be served by de-DE. "exact" requires the whole tag.

Inherited from

SpeechCommonOptions.localeMatch

---

mediaTypes?

optional mediaTypes?: string[]

Acceptable media types; always a hard requirement. Absent lets the provider choose.

---

model?

optional model?: string

Model profile to route within, as configured on the engine. Omitted or "auto" routes across every registered target. Never a vendor model or voice id.

Inherited from

SpeechCommonOptions.model

---

requestId?

optional requestId?: string

Fixes the run id, for correlation.

Inherited from

SpeechCommonOptions.requestId

---

telemetry?

optional telemetry?: Telemetry

Inherited from

SpeechCommonOptions.telemetry

---

text

text: string

The text to speak. Plain text; markup such as SSML is out of scope.

Inherited from

SpeechCommonOptions.text

---

voice?

optional voice?: SpeechVoice | SpeechVoiceInfo & Pick<SpeechVoice, "requirement">

The voice: an entry from listVoices, or provider-neutral traits. A voice left "preferred" falls back to its traits when it is gone or another provider serves the run; a "required" one never relaxes silently.

Inherited from

SpeechCommonOptions.voice