More than 90 languages with auto-detection
With Yandex Translate you can translate between a wide array of languages (more than 90 are already available). Your service may send a request without a source language code: our API will detect it automatically.
There are two main approaches to machine translation in use today. The statistical approach splits text into short fragments, while the neural approach does not. This makes translation of individual words more accurate using the statistical approach. However, to analyze whole sentences and understand phrases within the context, you need machine intelligence based on neural networks. Our API employs both models, but the algorithm decides which model to apply in each case.
Adaptive algorithm continues learning
Machine learning has not only helped us to launch Yandex Translate, but also contributes to improving the quality of translation. More and more new texts are continuously processed between numerous language pairs. However, the data you send to the service is not used for such purposes.
Services that involve cross-cultural communication should include tools to facilitate understanding texts in different languages. This may include communication between different users or audiences, like employee-to-customer, agent-to-advertiser, etc. When messages are translated by Yandex Translate, everyone can read the chat or message thread in their own language.
Create your website in one language, and then immediately add versions for more languages. Make your products or services available to multinational audiences country-wide or world-wide.
Help customers who publish text content on your service to attract more loyal followers. For example, short blog posts can be released in Russian and English in parallel, without the need to translate the content manually.
Companies offering translation and localization services can substantially reduce the employee effort by pre-translating texts with Yandex Translate.