This quickstart guide shows you how to get started with the Azure Content Moderator API to moderate text for potentially offensive content, profanity, and personally identifiable information (PII).

Prerequisites

Before you begin, make sure you have the following:

• An Azure subscription. If you don't have one, create a free account.
• A Content Moderator resource created in the Azure portal. For instructions, see Create a Content Moderator resource.
• The subscription key and endpoint for your Content Moderator resource.
• A development environment set up with a programming language of your choice (e.g., Python, C#, Node.js).

Step 1: Create a Content Moderator Resource

If you haven't already, create a Content Moderator resource in the Azure portal. Navigate to Create a resource, search for "Content Moderator", and follow the prompts to create your resource. Note down the Key and Endpoint from the resource's overview page.

Step 2: Install the SDK

You can interact with the Content Moderator API using REST calls, but using the official SDKs is often more convenient. Install the appropriate SDK for your development language.

Python

pip install azure-cognitiveservices-vision-contentmoderator

C#

dotnet add package Azure.CognitiveServices.Vision.ContentModerator

Node.js

npm install @azure/cognitiveservices-content-moderator

Step 3: Write the Code

Below are examples for moderating text using Python, C#, and Node.js.

Python Example

This example demonstrates how to detect moderate and review text.

from azure.cognitiveservices.content_moderator import ContentModeratorClient
from msrest.authentication import CognitiveServicesCredentials

# Replace with your key and endpoint
subscription_key = "YOUR_SUBSCRIPTION_KEY"
endpoint = "YOUR_ENDPOINT"

client = ContentModeratorClient(endpoint, CognitiveServicesCredentials(subscription_key))

text_to_moderate = "This is a sample text that might contain offensive language like hell or damn."

try:
    response = client.text_moderation.analyze_text(
        locale="en",
        text_content=text_to_moderate,
        classify_terminals=True
    )

    print("Text Moderation Results:")
    if response.is_moderation_required:
        print(f"  Moderation required: {response.is_moderation_required}")
        print(f"  Profanity terms detected: {len(response.terms)} - {response.terms}")
        print(f"  Personal Identifiable Information (PII) detected: {response.pii}")
        print(f"  Category scores: {response.categories}")
    else:
        print("  No offensive content detected.")

except Exception as e:
    print(f"An error occurred: {e}")

Note: Remember to replace YOUR_SUBSCRIPTION_KEY and YOUR_ENDPOINT with your actual credentials.

Step 4: Run the Application

Execute your code. The output will show whether the text was flagged for moderation, list any detected profanities, and provide PII detection information.

Related topics

• Azure Cognitive Services Overview
• Content Moderator API Reference
• Image Moderation
• Video Moderation