The preferred model to handle translations in your app.
This determines which model your app uses for translating content. The framework uses this strategy
when available, or automatically selects an appropriate alternative based on device capabilities and
language availability.
Members (2)
get · iOS 26.4+
lowLatency
TypeScript
static get lowLatency(): TranslationSession_Strategy
Swift declaration
static let lowLatency: TranslationSession.Strategy
A translation strategy that provides fast translations using traditional models.
Use this strategy for scenarios that require low latency, such as translating audio in real time.
This strategy requires downloading languages before use, but after they are downloaded they
are available to all apps on the device. To download languages ahead of time, call
``TranslationSession/prepareTranslation()`` before performing translations.
Compared to ``highFidelity``, this strategy is faster and uses less power, though
translations are not as fluent. This is the default strategy for devices without
Apple Intelligence and apps built with SDKs before iOS 26.4 and macOS 26.4.
get · iOS 26.4+
highFidelity
TypeScript
static get highFidelity(): TranslationSession_Strategy
Swift declaration
static let highFidelity: TranslationSession.Strategy
A translation strategy that provides more fluent translations using Apple Intelligence.
Use this strategy for higher-quality translations on devices with Apple Intelligence enabled.
The models are already downloaded when Apple Intelligence is enabled, so no additional
language downloads are required. This strategy offers higher-quality translations and
supports additional languages, but may take longer to complete than ``lowLatency``.
On devices without Apple Intelligence, it falls back to the traditional models used by
`lowLatency`.
