> ## Documentation Index
> Fetch the complete documentation index at: https://www.cometchat.com/docs/llms.txt
> Use this file to discover all available pages before exploring further.

# OpenAI Moderation

> OpenAI Moderation — CometChat documentation.

CometChat allows you to integrate OpenAI for real-time message moderation, enabling automated detection of harmful, offensive, or inappropriate content.

## Integration Steps

### Step 1: Create an OpenAI Prompt List

First, you need to create a list that defines your moderation prompt.

1. Navigate to the [CometChat Dashboard](https://app.cometchat.com) and select your app.

2. Go to **Moderation** → **Settings** from the left-hand menu.

3. Click on the **List** tab.

4. Click **Add List** to create a new list.

5. Fill in the required fields:
   * **Name**: Enter a descriptive name for your OpenAI moderation list (e.g., `openai-moderation`)
   * **ID**: Enter a unique identifier for the list (e.g., `openai-moderation-list`)
   * **Category**: Select **OpenAI Prompt** from the dropdown
   * **Description**: Enter a description (optional)
   * **Prompt**: Enter your moderation prompt. This prompt defines the criteria OpenAI will use to evaluate content. For example, to detect phone numbers:
     ```
     Detect if the message contains any phone numbers, mobile numbers, or contact numbers in any format.
     ```

6. Click **Save** to create the list.

### Step 2: Configure Advanced Settings

Next, configure the OpenAI model and error handling settings for your OpenAI API.

1. Navigate to **Moderation** → **Settings**.

2. Click on the **Advanced Settings** tab.

3. Scroll down to the **OpenAI** section.

4. Configure the following options:

   * **Model**: Enter the OpenAI model you want to use (e.g., `gpt-4-turbo`)

   * **OpenAI Key**: Enter your OpenAI API key to authenticate requests

   * **The number of messages sent to OpenAI for context**:
     * Set the number of previous messages to include as context (0-10)
     * This enables context-aware moderation by providing conversation history
     * Set to `0` if you don't need context and only want to moderate individual messages

   * **On OpenAI Error**:
     * **Approve Message** (Default): Messages are automatically approved if OpenAI is unavailable
     * **Block Message**: Messages are blocked when OpenAI is inaccessible, ensuring no unmoderated content passes through

5. Click **Save** to apply the settings.

<Frame>
  <img src="https://mintcdn.com/cometchat-22654f5b/M8Nbc5hMBI_9xvrR/images/moderation-openai-advanced-settings.png?fit=max&auto=format&n=M8Nbc5hMBI_9xvrR&q=85&s=37701831ceee6aac9f0d40e5e6b45e45" alt="OpenAI Advanced Settings" width="2534" height="716" data-path="images/moderation-openai-advanced-settings.png" />
</Frame>

### Step 3: Create a Moderation Rule

Now, map your OpenAI list to a moderation rule to activate the moderation.

1. Navigate to **Moderation** → **Rules**.

2. Click **Create New Rule**.

3. Fill in the mandatory fields:
   * **Name**: Enter a name for the rule (e.g., `test-openai`)
   * **ID**: Enter a unique identifier (e.g., `testopenai`)
   * **Description**: Optionally describe what this rule does

4. Configure the **Conditions**:
   * Select the content type to moderate (e.g., **Text**)
   * Select **Contains**
   * Select **OpenAI**
   * Choose the OpenAI list you created in Step 1 from the dropdown
   * Set the **confidence** threshold (e.g., `greater than 80%`)
     * Higher percentages result in stricter matching
     * Lower percentages allow more content to pass through

5. Configure the **Actions** to perform when content is flagged (block, flag for review, etc.).

6. Click **Save** to create the rule.

7. **Enable the rule** by toggling it on.

<Frame>
  <img src="https://mintcdn.com/cometchat-22654f5b/M8Nbc5hMBI_9xvrR/images/moderation-openai-create-rule.png?fit=max&auto=format&n=M8Nbc5hMBI_9xvrR&q=85&s=79c5b8028fca4271a3ad00d6e6a4b33c" alt="Create OpenAI Rule" width="2486" height="1120" data-path="images/moderation-openai-create-rule.png" />
</Frame>

<Note>
  The rule you create should be of type **Text Contains** or **Image Contains** to work with OpenAI moderation.
</Note>

***