Trask NLP

It's modular structure provides you with a powerful solution that can be applied to many different tasks, which domain is primarily unstructured or structured text data, and is capable of automating lots of manual routines, such as document classification, document content analysis, keywords extraction, annotation, segmentation/clusterization and entities extraction. However, the spectrum of NLP applications is infinite and our solution can be adjusted to your requirements.

Trask NLP API had initially been designed to satisfy chatbots' language needs (recognize user's intent, find entities and patterns in text, etc.) and consequently was extended to general-purpose NLP. Hence, it can be utilised for both chatbots and other domains.

  Obviously, the essential part of the API is its infrastructure: asynchronous I/O operations and hard calculations with the possibility to check its status, Open API (2.0) Swagger (REST API) interface, clients' models isolation, API authorisation, modular logging, a combination of two powerful ML/DS languages (Python and R).