Azure AI Content Understanding (Preview)

Azure AI Content Understanding extracts content and fields from multimodal input including documents, images, audio, and video using AI-powered analysis.

This connector is available in the following products and regions:

Service Class Regions
Copilot Studio Standard All Power Automate regions except the following:
     -   US Government (GCC)
     -   US Government (GCC High)
     -   China Cloud operated by 21Vianet
     -   US Department of Defense (DoD)
Logic Apps Standard All Logic Apps regions except the following:
     -   Azure Government regions
     -   Azure China regions
     -   US Department of Defense (DoD)
Power Apps Standard All Power Apps regions except the following:
     -   US Government (GCC)
     -   US Government (GCC High)
     -   China Cloud operated by 21Vianet
     -   US Department of Defense (DoD)
Power Automate Standard All Power Automate regions except the following:
     -   US Government (GCC)
     -   US Government (GCC High)
     -   China Cloud operated by 21Vianet
     -   US Department of Defense (DoD)
Contact
Name Microsoft
URL https://aka.ms/content-understanding
Email contentunderstanding@microsoft.com
Connector Metadata
Publisher Microsoft
Website https://aka.ms/content-understanding
Privacy policy https://privacy.microsoft.com
Categories AI

Creating a connection

The connector supports the following authentication types:

API Key Provide Content Understanding resource name and API Key. All regions Shareable
Logic Apps Managed Identity Create a connection using a LogicApps Managed Identity LOGICAPPS only Shareable
Microsoft Entra ID Integrated Use Microsoft Entra ID to access All regions Not shareable
Default [DEPRECATED] This option is only for older connections without an explicit authentication type, and is only provided for backward compatibility. All regions Not shareable

API Key

Auth ID: keyBasedAuth

Applicable: All regions

Provide Content Understanding resource name and API Key.

This is shareable connection. If the power app is shared with another user, connection is shared as well. For more information, please see the Connectors overview for canvas apps - Power Apps | Microsoft Docs

Name Type Description Required
API Key securestring Content Understanding API Key True
Endpoint URL string Content Understanding Endpoint Url (Example: https://your-resource.cognitiveservices.azure.com/) True

Logic Apps Managed Identity

Auth ID: managedIdentityAuth

Applicable: LOGICAPPS only

Create a connection using a LogicApps Managed Identity

This is shareable connection. If the power app is shared with another user, connection is shared as well. For more information, please see the Connectors overview for canvas apps - Power Apps | Microsoft Docs

Name Type Description Required
LogicApps Managed Identity managedIdentity Sign in with a Logic Apps Managed Identity True
Endpoint URL string Content Understanding Endpoint Url (Example: https://your-resource.cognitiveservices.azure.com/) True

Microsoft Entra ID Integrated

Auth ID: oauth

Applicable: All regions

Use Microsoft Entra ID to access

This is not shareable connection. If the power app is shared with another user, another user will be prompted to create new connection explicitly.

Name Type Description Required
Endpoint URL string Content Understanding Endpoint Url (Example: https://your-resource.cognitiveservices.azure.com/) True

Default [DEPRECATED]

Applicable: All regions

This option is only for older connections without an explicit authentication type, and is only provided for backward compatibility.

This is not shareable connection. If the power app is shared with another user, another user will be prompted to create new connection explicitly.

Name Type Description Required
API Key securestring Content Understanding API Key True
Endpoint URL string Content Understanding Endpoint Url (Example: https://your-resource.cognitiveservices.azure.com/) True

Throttling Limits

Name Calls Renewal Period
API calls per connection 1200 60 seconds

Actions

Analyze Content

Extract content and fields from multimodal input (documents, images, audio, video) using an analyzer. The operation runs asynchronously and returns the result when complete.

Analyze Content

Extract content and fields from multimodal input (documents, images, audio, video) using an analyzer. The operation runs asynchronously and returns the result when complete.

Parameters

Name Key Required Type Description
Analyzer ID
analyzerId True string

The unique identifier of the analyzer to use for analysis.

File Content
inputFileContent byte

File content to analyze (document, image, audio, or video). Base64-encoded. Cannot be combined with Input File URL or inputs.

Input File URL
inputFileUrl string

URL of the file to analyze. Cannot be combined with File Content or inputs.

Range
range string

Page range to analyze (e.g. 1-3,5,9-). Only applies when using File Content or Input File URL. If not set, all pages are processed.

File MIME Type
inputFileMimeType string

MIME type of the file content (e.g. application/pdf, image/jpeg). Optional when using File Content.

url
url string

URL of the input to analyze. Only one of url or data should be specified.

data
data byte

Base64-encoded binary content of the input to analyze. Only one of url or data should be specified.

name
name string

Name of the input.

MIME Type
mimeType string

The MIME type of the input content. Ex. application/pdf, image/jpeg, etc.

range
range string

Range of the input to analyze (ex. 1-3,5,9-). Document content uses 1-based page numbers.

Processing Location
processingLocation string

The location where the data may be processed. Defaults to global.

String Encoding
stringEncoding string

The string encoding format for content spans in the response. Possible values are 'codePoint', 'utf16', and utf8.

Returns

Status and result of the analyze operation.

Definitions

OperationState

Status of the operation.

Status of the operation.

AnalysisContent

Extracted content from analysis.

Name Path Type Description
kind
kind string

Content kind.

markdown
markdown string

Markdown representation of the content.

fields
fields object

Extracted fields from the content.

category
category string

Classified content category.

analyzerId
analyzerId string

The analyzer that generated this content.

mimeType
mimeType string

The MIME type of the content.

path
path string

The path of the content in the input.

AnalysisResult

Analyze operation result.

Name Path Type Description
analyzerId
analyzerId string

The unique identifier of the analyzer.

apiVersion
apiVersion string

The version of the API used to analyze the document.

createdAt
createdAt date-time

The date and time when the result was created.

warnings
warnings array of object

Warnings encountered while analyzing the document.

code
warnings.code string
message
warnings.message string
Markdown Content
_markdown string

Markdown representation of the first content item, for quick access.

contents
contents array of AnalysisContent

The extracted content.

AnalyzeOperationResult

Status and result of the analyze operation.

Name Path Type Description
id
id string

The unique ID of the operation.

status
status OperationState

Status of the operation.

code
error.code string

Error code.

message
error.message string

Error message.

result
result AnalysisResult

Analyze operation result.