If the request was successful, this will contain the results. For CreateSpeech requests, an audio file in a base 64 encoded string and for SpeechToText requests, the text that was transcribed.
OptionaldataOptionalerrorHow a failure is classified, when the request failed: ErrorAnalyzer's reading of the error
the provider's SDK threw, which keeps its HTTP status. A caller deciding whether to try another
vendor reads this, as the failover loop reads BaseResult.errorInfo: a rate limit or an outage
is worth another vendor, and a request the vendor rejected as invalid is not. The message alone
cannot tell them apart. Undefined on success, and from a driver that reports only
errorMessage.
OptionalerrorIf the request failed, this will contain the error message
True if the request was successful, false otherwise
OptionalusageUsage for the request, when the provider reported enough to build it.
Audio models are billed by duration rather than by token, so this is normally a
ModelUsage.ForMedia instance in Seconds — the quantity a per-minute or per-hour
price unit type prices. Left undefined when the provider did not report a duration, so that
cost calculation declines rather than billing the request as free.
Return payload for each CreateSpeech request