Technical overview of the ModelFront API

Read more about risk prediction and its use cases.

If you are a current client with an urgent issue or question, email [email protected] now.

Getting started

To get started with ModelFront, sign up to get an API key and access to the sandbox. Then you're ready to play with it and hit our REST API.


The risk API provides real-time risk predictions on each machine translation. You get a risk prediction score between 0.0 and 1.0 on each machine translation line, typically a sentence or a phrase.

You set your own quality threshold to control the balance of quality, scale and speed for each type of text.

If you do not include a translation in a request, a translation will be automatically pulled from the default engine for the language pair. The default engine is DeepL for the language pairs that it supports, and otherwise Google Translate.

Error types

ModelFront is built to catch many kinds of errors, both the common ones and the rare but painful ones.

Named entities

Errors translating names of persons, places, organisation and products


Revered meaning due to missing or inserted negation


Offensive or ambiguous translations


Untranslated words or phrases


Garbage input or output characters, words or phrases


Many other types of translation errors, for example translating from the wrong language

Have another type of error you want to catch? Let us know!


ModelFront is built to support any language pair with a sizable open corpus and supported by the Google Translate API and the Microsoft Translator API, including most pairs supported by the the Yandex Translate API, the IBM Watson Translator API, the Baidu Translate API and the DeepL API.

Afrikaans, Albanian, Arabic, Armenian, Azerbaijani (Latin), Basque, Belarusian, Bengali, Bosnian, Bulgarian, Burmese, Catalan, Cebuano, Chechen, Chinese (Simplified), Chinese (Traditional), Croatian, Czech, Danish, Dutch, English, Estonian, Finnish, French, Galician, Georgian, German, Greek, Gujarati, Haitian, Hebrew, Hindi, Hungarian, Icelandic, Indonesian, Irish, Italian, Japanese, Javanese, Kannada, Kazakh, Kyrgyz, Korean, Latin, Latvian, Lithuanian, Luxembourgish, Macedonian, Malagasy, Malay, Malayalam, Marathi, Mongolian, Nepali, Norwegian (Bokmal), Norwegian (Nynorsk), Persian, Polish, Portuguese, Punjabi, Romanian, Russian, Scots, Serbian, Serbo-Croatian, Slovak, Slovenian, Spanish, Sundanese, Swahili, Swedish, Tagalog, Tajik, Tamil, Telugu, Thai, Turkish, Ukrainian, Urdu, Uzbek, Vietnamese, Welsh, Yoruba


ModelFront works with any translation API. It is trained and tested against the major ones - Google Translate, DeepL, Microsoft Translator, Yandex.Translate (Яндекс), IBM Watson Language Translator, Baidu Fanyi (百度), Systran Translation and Amazon Translate.

It also works with custom-trained models like those from Google AutoML and Microsoft Custom Translator, which are key to many of the applications successfully using machine translation at scale.


ModelFront invests in custom development like crawling, labelling and training to hit your accuracy goals.

You can also add your own monolingual data, parallel data or labelled data to boost accuracy quickly and accelerate the rate at which ModelFront improves on your data.

ModelFront continues to automatically improve as you use it.


The default plans are annual and subscription-based. Volume-based plans are also available.

Security and Privacy

We can silo your proprietary data. To discuss dedicated hosting and on-premise deployment, please contact us.