In a nutshell

The Cortical.io Semantic Search Engine is a search system that understands natural language. It picks out precise, relevant information buried in terabytes of text repositories and instantly puts it at your fingertips. Our engine uses a unique meaning-based algorithm to efficiently resolve ambiguity and vocabulary mismatch issues. By measuring the semantic overlap, the engine is able to match a query with indexed information even if they do not use the same words. For example, the engine recognizes that the phrases “problems with sheets getting stuck in the printer” and “solving a paper jam error” are semantically similar.

With the Cortical.io Semantic Search Engine, companies are dramatically improving the efficiency of their search processes by increasing the accuracy and relevancy of search results. This leads to both increased customer and employee satisfaction and lower processing costs.

What makes Cortical.io Semantic Search different?

  • Represents every word with roughly 16,000 semantic features, which allows for very fine semantic distinctions
  • Requires little training material, which is particularly helpful in use cases where such material is scarce
  • Takes only a few hours to index your entire repositories
  • Is customized to your use case and fully operable within a few days

One search engine, many use cases

How Cortical.io’s Semantic Search Engine is helping a global company automate its marketing-claim verifications

view case study

Semantic Search FAQs

  • What kind of information does the Cortical.io Semantic Search Engine process?
  • How does the engine handle ambiguous search queries?
  • Does the engine understand long sentences?
  • How long does the training take?
  • How long does it take before I get a working Semantic Search Engine?
get the answers

Setting up: indexing your company’s information

  • The documents are cleaned and sliced into meaningful sections
  • The meaning of each section is numerically encoded as a semantic fingerprint
  • An inverted index of section fingerprints is created
  • You can easily add, change, or delete document sections in the index

Using: searching for information

  • You enter a search query that consists of words, semantic expressions, text, or sample documents
  • The system numerically encodes the meaning of the search query as a semantic fingerprint
  • The fingerprint of the search query is instantly compared with the fingerprints of the indexed documents
  • Results are ranked based on semantic similarity
  • Accessible through a REST API