TokenMill is a linguistic engineering company focusing on providing text analytics and search solutions. Founded in November 2011 TokenMill has worked on a number of government and commercial projects. Our services are falling into three main areas: Natural language proceeding. One common NLP task is identification of various entities of interest within unstructured text: organisation names, disease symptoms, product features and a like. The other is various techniques of document grouping: finding similar documents, near duplicates, themes within the stream of text, etc. Crawling. To do NLP one needs to get to the data first. Textual data is usually acquired via web or intranet crawling. We do have the expertise needed to run large scale and focused crawls. Vertical search. Lastly all the collected and analysed data has to be accessed by the end users. Usually, it is done via specialised search engines which allow to run custom queries, rank documents according to business domain rules and provide other tools allowing to query and reason with otherwise inaccessible huge textual data. We provide solutions in those areas partially based on internally developed tools - which in turn are based on solid open source projects - as well as on custom development tailored for the needs of a particular client.