About                     Contact                     Copyright © 2015 - 2019 DeepSearch

About DeepSearch

DeepSearch is a hybrid information retrieval system that merges search engine technology with deep learning.

Keyword search is supported through an automated query extension and an online autosuggest service, leveraging DeepSearch's proprietary machine learning technology.
In contrast to a traditional search solution which limits search queries to a few words, DeepSearch also supports queries made of entire documents.
This effectively makes DeepSearch a concept search engine enabling - for example - patent attorneys, academics, legal professionals or industry analysts to carry out a search with a patent specification, a research paper, a court’s jurisprudence or a press article.

Unlike the Semantic Web which requires the development of ontologies that are especially time consuming to produce, DeepSearch is trained automatically in a matter of hours - if not minutes - in any language or nomenclature and with any corpus.

Finally, there is the added convenience of allowing queries to be saved within the user’s browser by clicking on the link bellow the search box: no user information is stored on the server, DeepSearch being a stateless - as in no tracking - web service.


Document search takes at most a few milliseconds to produce sorted results from millions of indexed documents. However, to guarantee the latest search results Wiki DeepSearch calls the mediawiki web service for query search and article snippets, thus incurring a slight delay.
Also note that this Wikipedia instance of DeepSearch is a demonstrator model, consequently document search has been disabled (though similar article search is supported).