Language identification and content routing

Semantria can process content in many different languages. Each Semantria configuration has a language associated with it, and will only give meaningful results if the content submitted matches the language of the configuration.

Semantria can detect the content language but does not route content based on the detected language unless you submit documents to a configuration route. This is because you can create many configurations for the same language to support your needs and Semantria does not know which of those are best suited to the content you submit. However, for configuration routes, a configuration can be specified for each language.

If you submit documents to a configuration route without specifying a language ID then language identification will be run on each document and the document will be submitted to the configuration you associated with that language for the route. If no configuration has been specified for the language that was detected, the document will be processed with the default configuration. Whatever configuration was actually used in processing the document will be reported in the results.

You can control whether you want to see the detected language in the returned data at a configuration level. This can be useful if you have content where you do not know the language of the individual pieces, but will cost you API credits as any other content submission does.