OpenAI AddOn

1. Overview
2. General Settings
3. Monitor Usage
4. Display Usage Widget
5. Disable Prompts Per User
6. Disable AddOn Per User
7. Disable AddOn Per Content Block Type
8. Prompt Suggestions
9. Customize Prompt Suggestions
10. References

Overview

The OpenAI AddOn is fully functional from the moment it’s enabled via the Beefree SDK Console. For information about enabling OpenAI AddOn click here.

Once enabled, the OpenAI AddOn is available in the side panel for the following content blocks:

• Title
• Paragraph
• List
• Button

In certain scenarios, you may find the need to personalize both the user interface (UI) and the operational features of the OpenAI AddOn. This is particularly applicable when you want to achieve objectives such as:

1. Monitoring Token Usage for Cost Management: By tracking the number of tokens being used, you can effectively manage and regulate your expenses related to using the OpenAI AddOn. This becomes important when your usage is high, and you must keep a budget check. Learn more about tokens.
2. Controlling Access to the Prompt: You might want to limit who can access and manipulate the prompt. In a shared environment, for instance, you may want to restrict certain users from altering the prompt, which is crucial for maintaining consistency and avoiding unwanted changes.
3. Enabling Per User or Per Content Type: You might want to limit who can access the AI so you can up-sell the feature to end-users. Or, you may only want to enable the AI capability for specific content types, such as paragraphs vs. buttons.
4. Disabling Automatic Suggestions: The OpenAI AddOn can generate automatic suggestions at the prompt, which may not always be desirable. In such cases, you might want to disable this feature to have more control over the input and output at the prompt.

We’ll cover all of these scenarios in the following sections as we discuss the settings and configuration options available for developers.

General Settings

Note: All settings are optional and may be refreshed, in real-time, via the configuration reload instance method (e.g. bee.loadConfig(settings)) during the active session.

Currently, the following settings are supported:

Example

const beeConfig = {
    uid: 'string',
    ... 
    addOns: [
      {
        id: "ai-integration",
        settings: {
          tokensAvailable: tokensAvailable,
          tokensUsed: tokenCounter,
          tokenLabel: 'tokens',
          isPromptDisabled: false,
          isSuggestionsDisabled: false,
        }
      },
    ],
    ...
}
     

Monitor Usage

With each prompt response, the addon will report the usage data provided by the OpenAI API via the editor’s onInfo callback without storing or tracking the data.

Example

const beeConfig = {
    uid: 'string',
    ... 
    onInfo: function (data) {

        // Check information is addon
        if (data.code === 1000) {

            // Check which addon
            const handle = data.detail.handle
            if (handle === 'ai-integration') {
                // Get the total tokens
                const totalTokens = data.detail.usage.total_tokens
                // Update running total in database
                // --> host app code goes here
            }
        }
    },
    ...
}
     

Display Usage Widget

You may choose to track the end-user’s total usage through the aforementioned onInfo callback if desired. Additionally, you may choose to show the usage data to the user via the built-in display widget. To activate the display usage widget, provide the usage data via the addon settings.  Since the editor doesn’t track usage, you’ll need to refresh the values via the bee.loadConfig method to keep the display widget data current.

Example

let tokenCounter = 0
let tokensAvailable = 2000

const beeConfig = {
    uid: 'string',
    ...
    // Initial AddOn configuration
    addOns: [
      {
        id: "ai-integration",
        settings: {
          tokensAvailable: tokensAvailable,
          tokensUsed: tokenCounter,
          tokenLabel: 'words',
          isPromptDisabled: false,
        }
      },
    ],
    ...
    // Monitor for information
    onInfo: function (data) {

        // Check information code is 1000 for addon reporting
        if (data.code === 1000) {

            // Check which addon
            const handle = data.detail.handle
            if (handle === 'ai-integration') {
                const totalTokens = data.detail.usage.total_tokens
                tokenCounter = tokenCounter + totalTokens

                // Refresh Usage AddOn Settings
                const refreshedUsageSettings = {
                    addOns: [
                        {
                            id: "ai-integration",
                            settings: {
                                tokensAvailable: tokensAvailable,
                                tokensUsed: tokenCounter,
                                tokenLabel: 'tokens',
                                isPromptDisabled: (tokenCounter >= tokensAvailable) ? true : false,
                            }
                        },
                    ],
                }

                // Reload Config
                bee.loadConfig(refreshedUsageSettings)
            }
        }
    },
}

See below for an example of how the UI will render when provided with the optional display usage widget.

Disable Prompts Per User

To enable OpenAI AddOn, but disable the prompt per user, pass the isPromptDisabled boolean parameter as true.

The following example will disable the prompts for the user with an uid of inactive-user.

Example

const beeConfig = {
    uid: 'inactive-user',
    ... 
    addOns: [
      {
        id: "ai-integration",
        settings: {
          isPromptDisabled: true,
        }
      },
    ],
    ...
}
     

Disable AddOn Per User

const beeConfig = {
    uid: 'inactive-user',
    ... 
    addOns: [
      {
        id: "ai-integration",
        enabled: false
      },
    ],
    ...
}
     

Disable AddOn Per Content Block Type

The OpenAI AddOn is available for the following content blocks:

• Title
• Paragraph
• List
• Button

You may utilize the Advanced Permissions configuration to disable OpenAI AddOn per content type.

The following example will disable the addon for the paragraph block:

Example

const beeConfig = {
    uid: 'string',
    ... 
    advancedPermissions: {
      content: {
        paragraph: {
          behaviors: {
            canViewSidebar: true,
          },
          properties: {
            aiIntegration: {
              show: false,
              locked: true
            },
          },
        },
      },
    }
    ...
}
     

Prompt Suggestions

OpenAI AddOn includes preset prompt suggestions to facilitate the content creation process. These appear after the initial draft of your text has been formulated and whenever further refinement is needed. Please note that this function applies only when editing existing text through the AI prompt. The suggestions will not appear for placeholder text.

Here’s a simplified step-by-step guide on how to use suggestions, as shown below:

1. Start with your draft text added to the design area.
2. If you desire to adjust the tone to be more formal, for example, click on the paragraph, list, title, or button you want to modify.
3. Navigate to the ‘Write with AI’ option.
4. Click on the area designated for adding prompts.
5. You’ll notice the prompt suggestions popup.
6. Select the ‘Make it [desired tone]’ option, making sure to replace [desired tone] with a value (e.g., funny)
7. Click ‘Generate’.

The AI will then generate a revised version of your content, matching the tone you entered.

Customize Prompt Suggestions

Below are the preset prompt suggestions we have identified for the different content tiles, along with their corresponding translation key, incase you’d like to revise the prompt through our custom languages feature.

Paragraph

Button

List

Title

References

• OnError, onWarning, and onInfo Callbacks
• Custom Languages
• Advanced Permissions