Textual Content Augmentation
Keyword Extraction

Keyword Extraction

Keyword extraction enables you to extract key terms and entities from your text data.

The Named Entity Recognition (NER) module classifies tokens in text, identifying entities such as names, locations, and organizations.

There are two types of keyword extraction available as given below.


  • Usage: You can manually trigger keyword extraction through API requests, specifying the tokens filter in the _additional field.
  • Benefit: It provides real-time extraction, allowing for flexible application as needed.


  • Usage: In this case, keywords are automatically extracted as content is ingested and synchronized, becoming immediately accessible in the tokens field.
  • Benefit: The benefits include streamlining data retrieval, and offering pre-extracted keywords for efficiency.
Model NameCase SensitiveTraining DatasetPrimary ApplicationLanguageDescription
dbmdz-bert-large-cased-finetuned-conll03-englishYesCoNLL-03Named Entity RecognitionEnglishA BERT model fine-tuned on the CoNLL-03 dataset for Named Entity Recognition. The model is case-sensitive, performing better with proper casing.
dslim-bert-base-NERNoUnknownNamed Entity RecognitionMultilingualA base BERT model optimized for Named Entity Recognition tasks across various languages.
davlan-bert-base-multilingual-cased-ner-hrlYesUnknownNamed Entity RecognitionMultilingualA multilingual BERT model fine-tuned for Named Entity Recognition with case sensitivity. Suitable for high-resource languages.