← Back to Discover

TextRazor

A natural language processing API that pulls entities, key phrases, topics, relations and word dependencies out of text in 19 languages, running in the cloud or on your own infrastructure.

Data analysisModel APIs
Visit website ↗

What it does

TextRazor is an API for analysing text. It extracts entities, disambiguates them and links them to a knowledge base, pulls out key phrases, and tags and classifies content by topic. The site says this works across 19 languages and can classify to the full IAB Content and IPTC Media Topic taxonomies. Deeper analysis returns relations, typed dependencies between words, and synonyms, which the site presents as the basis for context-aware semantic applications. It combines natural language processing techniques with a large knowledgebase of facts and is aimed at documents, tweets and web pages. You can try a demo on the site or sign up for a free API key. Documentation includes a REST API reference and Python, PHP and Java references, and the site says the infrastructure can run in the cloud or be self-hosted.

Capabilities

  • Entity extraction, disambiguation and linking
  • Keyphrase extraction
  • Topic tagging and classification, including IAB and IPTC taxonomies
  • Relations, typed dependencies and synonyms
  • Analysis in 19 languages
  • Cloud or self-hosted deployment

When to use it

  • You need structured entities and topics from articles, tweets or web pages
  • You want to classify content against IAB or IPTC taxonomies
  • You want an NLP API you can call from Python, PHP, Java or REST