DeepL was founded in 2009 in Germany as Linguee, an online dictionary, which set out to create a neural machine translation system that could produce [translations of a much higher quality than traditional statistical machine translation. I suspect is also used chatGPT4 behind to assist.
Supported languages can be found here
DeepL offers paid and free accounts, free accounts allowing 500,000 characters per month free.
While DeepL performs extended functions like full document translation, this implementation is limited to text at this time.
wappler-all-in-one-ai can be installed from:
The additional environment variable DEEPL_API_KEY must be added to your workflow settings
Select Deepl from the "wappler-all-in-one-ai" module.
Max tokens and temperature related inputs are then hidden and 3 new inputs are revealed.
Translate From:
This is the language in which the text currently is. It can be specifed or left as "Auto detect" whereupon DeepL will identify the language for you.
Translate to:
This is the target language
Formality:
This tells the translation engine to use a more or less formal model for translation if required otherwise it is translated in the same style as submitted. Not all languages support formality and if not supported the setting is ignored.
Schema
wappler-all-in-one-ai now returns an extended schema like this showing the language and the formality setting.
Language is returned as a comma separated pair comprising the language from and language to settings"
{"multiAI1":
{"result":"Bonjour le monde",
"service":"deepl",
"engine":"text",
"maxTokens":"",
"temperature":"",
"language":"default,fr",
"formality":"default"}}
Max tokens and temperature are not relevant to DeepL and are returned empty