[API][yes]
[Search][no]
[UI][partial]
Translation engines take text content and translate it to another language, ideally outputting it in the same format as the input.
Engine input options
Translation engines are unique in that they can translate many different input types. The input formats the engine will receive are dependent on the supportInputFormats key in the engine manifest.
There are 5 types of input formats a translation engine may support:
Translation engines can choose to support one or more of these input formats by specifying the proper mime types in their engine manifest's supportedInputFormats field. See the individual sub-pages for details on how to support each individual type.
As the engine developer, you may decide which formats to support and whether to support them in a single engine or separate ones.
Custom fields
Leveraging custom fields is particularly important for translation engines due to the many different languages and modes they can run in. Often times, translation engines will define custom fields for inputs such as:
- Input language
- Output language
- Input document type (e.g. plain, rich, extracted, recognized, transcript)
- Output document type (e.g. plain, rich, extracted, recognized, transcript)
The values submitted to these fields by users can then be accessed by the engine in the message payload it receives.
