Ireul Neural Information Extraction Engine features the unique ability of exhaustive Open Information Extraction (OIE). Ireul understands knowledge expressed through natural languages, and transforms it into structured data, which enables automatic knowledge base construction and numerous business intelligence applications.

Through multiple specifically designed neural networks and hypothesis searching algorithm, Ireul is able to utilize in-depth semantic features and facilitate multilingual transfer learning, and achieves satisfying outcomes on multiple languages with zero resource.

Unlike traditional information extraction systems, Ireul discards artificially-set rules, semantic role labeling, dependency parsing, and other techniques that constrain the system’s capability of generalization, which in result gives Ireul advantage in processing interwound information, long-term dependency, Winograd Schema, and other complicated situations.

Ramiel is a performance-driven full text search engine, featuring an original TLA+ verified indexing structure that guarantees extraordinary response speed and reliable result quality at any scale.

The succinct data structure makes full use of different storage media to cooperate with the customized machine learning based ranking algorithm, and enables token-level delta update and various early termination strategies based on probability.

Ramiel also natively supports distributed deployment with inbuilt protocols for sharding and replication.

Arael provides fundamental support to the analysis of natural language, with the ability to handle more than 170 different languages in the form of stream. For languages like Chinese, Japanese, Korean, Thai, etc., Arael is able to deal with overlapping ambiguity and out-of-vocabulary words/phrases using machine learning.

Not only can Arael segment basic graphemes, words/phrases, and sentences, Arael can also be configured with analyzers and filters to form customized workflows.

In addition, Arael implemented Unicode® Standard Annex #29 and Technical Standard #51, translation: Never get screwed by Emojis!

Matarael is a distributed crawling system that complies with standards, and is capable of crawling and analyzing data at the scale of the Internet.

Instead of using bloated headless browsers, we designed Matarael from square one following numerous standards of IETF RFC and W3C, resulting in highly efficient stream-based crawling and thorough support for advanced features such as JavaScript. We place significant importance on compliance, such that Matarael will follow robots.txt, X-Robots-Tag, and any Robots Exclusion Protocol variations.

Matarael also supports distributed Web Graph construction and Map-Reduce analysis on various storage backends.