Key Phrase Extraction cognitive skill

The Key Phrase Extraction skill evaluates unstructured text, and for each record, returns a list of key phrases. This skill uses the machine learning models provided by Text Analytics in Cognitive Services.

This capability is useful if you need to quickly identify the main talking points in the record. For example, given input text "The food was delicious and there were wonderful staff", the service returns "food" and "wonderful staff".


As you expand scope by increasing the frequency of processing, adding more documents, or adding more AI algorithms, you will need to attach a billable Cognitive Services resource. Charges accrue when calling APIs in Cognitive Services, and for image extraction as part of the document-cracking stage in Azure Cognitive Search. There are no charges for text extraction from documents.

Execution of built-in skills is charged at the existing Cognitive Services pay-as-you go price. Image extraction pricing is described on the Azure Cognitive Search pricing page.



Data limits

The maximum size of a record should be 50,000 characters as measured by String.Length. If you need to break up your data before sending it to the key phrase extractor, consider using the Text Split skill.

Skill parameters

Parameters are case-sensitive.

Inputs Description
defaultLanguageCode (Optional) The language code to apply to documents that don't specify language explicitly. If the default language code is not specified, English (en) will be used as the default language code.
See Full list of supported languages.
maxKeyPhraseCount (Optional) The maximum number of key phrases to produce.
modelVersion (Optional) The version of the model to use when calling the Text Analytics service. It will default to the latest available when not specified. We recommend you do not specify this value unless absolutely necessary. See Model versioning in the Text Analytics API for more details.

Skill inputs

Input Description
text The text to be analyzed.
languageCode A string indicating the language of the records. If this parameter is not specified, the default language code will be used to analyze the records.
See Full list of supported languages

Skill outputs

Output Description
keyPhrases A list of key phrases extracted from the input text. The key phrases are returned in order of importance.

Sample definition

Consider a SQL record that has the following fields:

    "content": "Glaciers are huge rivers of ice that ooze their way over land, powered by gravity and their own sheer weight. They accumulate ice from snowfall and lose it through melting. As global temperatures have risen, many of the world’s glaciers have already started to shrink and retreat. Continued warming could see many iconic landscapes – from the Canadian Rockies to the Mount Everest region of the Himalayas – lose almost all their glaciers by the end of the century.",
    "language": "en"

Then your skill definition may look like this:

    "@odata.type": "#Microsoft.Skills.Text.KeyPhraseExtractionSkill",
    "inputs": [
        "name": "text",
        "source": "/document/content"
        "name": "languageCode",
        "source": "/document/language" 
    "outputs": [
        "name": "keyPhrases",
        "targetName": "myKeyPhrases"

Sample output

For the example above, the output of your skill will be written to a new node in the enriched tree called "document/myKeyPhrases" since that is the targetName that we specified. If you don’t specify a targetName, then it would be "document/keyPhrases".


              "world’s glaciers", 
              "huge rivers of ice", 
              "Canadian Rockies", 
              "iconic landscapes",
              "Mount Everest region",
              "Continued warming"

You may use "document/myKeyPhrases" as input into other skills, or as a source of an output field mapping.


If you provide an unsupported language code, a warning is generated and key phrases are not extracted. If your text is empty, a warning will be produced. If your text is larger than 50,000 characters, only the first 50,000 characters will be analyzed and a warning will be issued.

See also