Named Entity Extraction — identify people and places

Extract Named Entities from Text API

What it does

Named Entity Extraction identifies people, organisations, locations, and miscellaneous entities in text, and returns each match with character offsets.

Send a text string in the request body, and you get the original text back along with an entities array. Each entity includes the extracted word, its type (PER, ORG, LOC, or MISC), start and end offsets, and a confidence score. The input text is limited to 512 tokens, so it is designed for short passages, snippets, and document segments rather than full-length files.

Use Named Entity Extraction when you need to index articles, route support tickets, enrich content metadata, or prepare text for downstream NLP steps. The offsets make it easier to highlight entities in a UI, map results back to source text, or combine extraction with other text-processing logic.

This endpoint is a good fit for applications that need lightweight entity recognition without building and maintaining their own NLP pipeline.

curl -X POST "http://localhost:8080/dosvak/extract-named-entities" \ -H "apy-token: $APY_TOKEN" \ -H "Content-Type: application/json" \ -d '{"text":"Barack Obama visited Microsoft in Seattle."}'

What you'll get back

Returns a JSON object with text as a string and entities as an array of entity objects.

Each entity object can include start and end integer offsets, type (PER, ORG, LOC, or MISC), word as the extracted text, and score as a number.

{
  "text": "Barack Obama visited Microsoft in Seattle.",
  "entities": [
    {
      "start": 0,
      "end": 12,
      "type": "PER",
      "word": "Barack Obama",
      "score": 0.99
    }
  ]
}

About this endpoint

What it does

Extracts named entities from the input text and returns the original text along with a list of detected entities, including their character offsets, type, surface form, and score.

Request Body

Parameter	Type	Mandatory	Description
text	String	Yes	Input text. Maximum length: 512 characters. Truncated to 512 tokens.

Response

Returns a JSON object with a text string field and an entities array field. Each item in entities is an object containing the detected entity’s offsets, label, text, and score.

Parameter	Type	Mandatory	Description
text	String	No	The input text as returned by the service.
entities	Object Array	No	The detected entities. Each item includes `start`, `end`, `type`, `word`, and `score`.
entities[].end	Integer	No	End offset of the entity in the text.
entities[].type	ENUM	No	Entity label. Allowed values: `PER`, `ORG`, `LOC`, `MISC`.
entities[].word	String	No	The extracted entity text.
entities[].score	Number	No	Confidence score for the entity.
entities[].start	Integer	No	Start offset of the entity in the text.

Body

Name

Type

Description

bodyREQUIRED

object

—

Extract Named Entities from Text API

What it does

Quickstart

What you'll get back

About this endpoint

What it does

Request Body

Response

Body

Errors any endpoint can return