Microsoft Azure Documentation

Text Analytics API Reference

This document provides detailed reference information for the Azure Text Analytics API. This service enables you to analyze text using natural language processing (NLP) capabilities.

Features Supported

  • Sentiment Analysis
  • Key Phrase Extraction
  • Named Entity Recognition (NER)
  • Language Detection
  • Personally Identifiable Information (PII) Detection
  • Healthcare Text Analytics (Preview) Preview

API Endpoints

POST /text/analytics/v3.0/sentiment

Analyzes the sentiment of a given text.

POST /text/analytics/v3.0/keyPhrases

Extracts key phrases from a given text.

POST /text/analytics/v3.0/entities

Recognizes entities in a given text.

POST /text/analytics/v3.0/languages

Detects the language of a given text.

POST /text/analytics/v3.0/pii

Detects Personally Identifiable Information (PII) in a given text.

POST /text/analytics/v3.1-preview.3/health

Analyzes healthcare-specific text. Preview

Authentication

Requests to the Text Analytics API are authenticated using a subscription key. You can obtain a key from the Azure portal.

Include your subscription key in the request header as follows:

Ocp-Apim-Subscription-Key: YOUR_SUBSCRIPTION_KEY

Request Body Structure

Most Text Analytics operations accept a JSON payload containing an array of documents to be processed. Each document requires a unique ID and the text content.


{
    "documents": [
        {
            "id": "1",
            "language": "en",
            "text": "This is a sample document."
        },
        {
            "id": "2",
            "language": "es",
            "text": "Este es otro documento de ejemplo."
        }
    ]
}

Common Parameters

Parameter Type Description Required
documents Array of Document A list of documents to be analyzed. Yes
id String Unique identifier for the document. Yes
language String The ISO 639-1 language code of the text. If not specified, the service will attempt to detect the language. No
text String The text content to analyze. Yes
stringIndexType String Specifies the index type used for string offsets. Can be 'TextElements' or 'UnicodeCodeUnit'. Defaults to 'UnicodeCodeUnit'. No

Example: Sentiment Analysis Request

Here's an example of sending a POST request to the sentiment analysis endpoint:


POST https://YOUR_TEXT_ANALYTICS_ENDPOINT/text/analytics/v3.0/sentiment?showStats=true
Content-Type: application/json
Ocp-Apim-Subscription-Key: YOUR_SUBSCRIPTION_KEY

{
    "documents": [
        {
            "id": "sentimentDoc1",
            "language": "en",
            "text": "The weather today is beautiful and sunny!"
        },
        {
            "id": "sentimentDoc2",
            "language": "en",
            "text": "I am not happy with the service I received."
        }
    ]
}

Response Structure

The API returns a JSON object containing the results for each document submitted, organized by operation.


{
    "sentimentDocuments": [
        {
            "id": "sentimentDoc1",
            "sentiment": "positive",
            "confidenceScores": {
                "positive": 0.98,
                "neutral": 0.01,
                "negative": 0.01
            },
            "warnings": []
        },
        {
            "id": "sentimentDoc2",
            "sentiment": "negative",
            "confidenceScores": {
                "positive": 0.05,
                "neutral": 0.10,
                "negative": 0.85
            },
            "warnings": []
        }
    ],
    "errors": [],
    "modelVersion": "2020-04-01"
}

Refer to the specific API operation documentation for detailed response structures.

Further Reading