Search and analyze Big Text Data repositories cost-efficiently

The Retina Library (formerly called Retina Spark) has been designed for high performance semantic text processing. It can classify, filter and search huge text repositories as well as streaming text (e.g. social media feeds) based on meaning, not keywords. The performance of the underlying algorithm enables to reduce the number of false positives drastically while increasing the percentage of recall.

Retina Library at a glance

  • Core NLU module:
    • Conversion of words or text items into semantic fingerprints
    • Computation of Boolean operations with semantic fingerprints (semantic expressions)
    • Text functions like segmentation, keyword extraction, parsing, identification of contexts and similar terms
    • Comparison of semantic fingerprints
    • Decoding context words of semantic fingerprints
    • Access to Retina Database General English
  • Semantic search module
  • Classifier module (dynamic generation of classifiers)
  • High performance semantic processing module (Spark)


  • Access via a Java and Scala API 
  • Integration in Apache Spark or in single-node Java runtime environments 
  • Usable with on premise or in the cloud Spark clusters (Cloudera, Amazon EMR) 
  • Cloudera certified 
  • Supports all Hadoop storage formats
  • Contact [email protected] for more information