npm.io
2.6.3 • Published yesterday

@iblai/agent-ai

Licence
MIT
Version
2.6.3
Deps
0
Size
240 kB
Vulns
0
Weekly
0

AgentAI Web Component

The AgentAI web component is a customizable chat widget that can be embedded in any web application. This component is distributed via npm and can also be accessed directly via a bundled JavaScript file hosted on an S3 bucket.

Note: The component is registered as both <agent-ai> and <mentor-ai> for backwards compatibility. <mentor-ai> will be removed in a future version — prefer using <agent-ai>.


Features

  • Embeddable chat widget.
  • Configurable via attributes.
  • Supports authenticated and anonymous modes.
  • Context-aware functionality for dynamic page data exchange.

Installation

Using npm
npm install @iblai/agent-ai
Direct Usage via S3

Include the bundled script in your HTML file:

<script src="https://assets.ibl.ai/web/mentorai.js?versionId=LaZKI7VC02L5buFhcIcaKiBKf4SyjTkD"></script>

Usage

Vanilla Javascript

Include the component directly in your HTML:

<!DOCTYPE html>
<html lang="en">
  <head>
    <meta charset="UTF-8" />
    <meta name="viewport" content="width=device-width, initial-scale=1.0" />
    <script src="https://assets.ibl.ai/web/mentorai.js"></script>
    <title>AgentAI Widget</title>
  </head>
  <body>
    <agent-ai
      mentorurl="https://mentor.iblai.app"
      authurl="https://auth.iblai.app"
      lmsurl="https://learn.iblai.app"
      tenant="your-tenant"
      mentor="mentor-id"
      isanonymous
      iscontextaware
      isadvanced
      theme="light"
    ></agent-ai>
  </body>
</html>
Attributes
Attribute Description Type Default Value
mentorurl URL of the MentorAI platform. String https://mentor.iblai.app
tenant Tenant name for authentication. String undefined
mentor Mentor name for the chat widget. String undefined
isanonymous Enables anonymous mode. If set, the component will not require user authentication. (Ensure this matches the anonymity of the mentor) Boolean false
iscontextaware Enables context-aware functionality to send page information to the iframe. Boolean false
isadvanced Enables advanced chat features. Boolean false
authUrl URL for authentication. String https://auth.iblai.app
contextOrigins Comma separated values indicating the origins whitelisted for sending context. Defaults to an empty string. String ""
lmsUrl URL for edX LMS. String learn.iblai.app
authRelyOnHost Used to determine if to solely depend on the host for authentication and prevent mentor iframe from redirecting to the auth SPA. Boolean false
edxUserId Indicates the user id on edX. Optional. String undefined
theme Sets the theme of the component. Can be either light or dark. String light
component Specifies the component to display in the mentor application. Can be one of analytics-overview, analytics-users, analytics-topics, or chat. String chat
redirectToken A token used by the auth layer to redirect back to the equivalent set URL. String undefined
enablechatactionpopup Enables chat action popups for voice calls and screen sharing. When enabled, sends MENTOR:ENABLE_CHAT_ACTION_POPUPS message to the iframe. Boolean false
contextsettings Enables the LMS context settings view. When present, renders a settings form instead of the chat iframe. Boolean false
contextid The LMS context/course identifier. Displayed in the heading and sent with the settings API request. Used with contextsettings. String undefined
contextenabled Sets the initial state of the Enabled checkbox in the context settings view. Present = checked, absent = unchecked. Used with contextsettings. Boolean false

LMS Context Settings View

When the contextsettings attribute is present, the component renders a settings form instead of the chat iframe. This view is designed for instructors in external LMS systems (e.g. Canvas, Open edX) who need to configure whether the mentor AI widget should be enabled or disabled for a specific course or context.

What the Settings Form Displays
  • A heading showing the Context ID (e.g. "Editing LMS Context Id: 110")
  • An Enabled checkbox to toggle whether the mentor widget is active for this context
  • A Mentor ID text input to specify which mentor to use
  • A Save button that submits the configuration via a POST request
How It Works
  1. The instructor sees a form pre-populated with the provided attribute values (enabled state, mentor ID).
  2. They can toggle the Enabled checkbox and/or change the Mentor ID.
  3. Clicking Save sends a POST request to:
    {lmsurl}/api/mentor-xblock/orgs/{tenant}/context/?context_id={contextid}
    with a JSON body:
    {
      "enabled": true,
      "mentor_id": "744a3672-0a59-4cc9-ad5a-03ada402a658"
    }
  4. On success (HTTP 200), a green "Settings saved successfully" message is displayed.
  5. On error, the component displays the error message from the response payload (e.g. {"error": "Mentor ID must be a valid UUID"}), or a generic "An unknown error occurred" message if no specific error is returned.
Required Attributes for Context Settings
Attribute Required Description
contextsettings Yes Boolean flag that enables the context settings view.
contextid Yes The LMS context/course identifier displayed in the heading.
lmsurl Yes Base URL of the LMS, used to construct the API endpoint.
tenant Yes Platform/organization key, used in the API endpoint path.
mentor Yes The initial Mentor UUID pre-filled in the Mentor ID input field.
contextenabled No If present, the Enabled checkbox is initially checked. Otherwise unchecked.
Context Settings Examples
Enabled context with a pre-filled mentor
<agent-ai
  contextsettings
  contextid="110"
  contextenabled
  lmsurl="https://learn.iblai.app"
  tenant="my-org"
  mentor="744a3672-0a59-4cc9-ad5a-03ada402a658"
></agent-ai>

This renders the form with the Enabled checkbox checked and the Mentor ID field pre-filled with the provided UUID.

Disabled context with no mentor
<agent-ai
  contextsettings
  contextid="205"
  lmsurl="https://learn.iblai.app"
  tenant="my-org"
  mentor=""
></agent-ai>

This renders the form with the Enabled checkbox unchecked and an empty Mentor ID field.

React usage
import React from "react";
import "@iblai/agent-ai";

const CourseSettings = ({ contextId, mentorId, isEnabled }) => {
  return (
    <agent-ai
      contextsettings
      contextid={contextId}
      {...(isEnabled ? { contextenabled: "" } : {})}
      lmsurl="https://learn.iblai.app"
      tenant="my-org"
      mentor={mentorId}
    ></agent-ai>
  );
};

Javascript Frameworks

React

Install the package:

npm install @iblai/agent-ai

Use the component

import React from "react";
import "@iblai/agent-ai";

const App = () => {
  return (
    <div>
      <agent-ai
        mentorurl="https://mentor.iblai.app"
        authurl="https://auth.iblai.app"
        lmsurl="https://learn.iblai.app"
        tenant="your-tenant"
        mentor="mentor-id"
        isanonymous
        iscontextaware
        isadvanced
        theme="light"
      ></agent-ai>
    </div>
  );
};

export default App;
Angular

Install the package:

npm install @iblai/agent-ai

Use the component in your template:

<agent-ai
  mentorurl="https://mentor.iblai.app"
  authurl="https://auth.iblai.app"
  lmsurl="https://learn.iblai.app"
  tenant="your-tenant"
  mentor="mentor-id"
  isanonymous
  iscontextaware
  isadvanced
  theme="light"
></agent-ai>

Add CUSTOM_ELEMENTS_SCHEMA in your module:

import { NgModule, CUSTOM_ELEMENTS_SCHEMA } from "@angular/core";

@NgModule({
  declarations: [AppComponent],
  imports: [BrowserModule],
  schemas: [CUSTOM_ELEMENTS_SCHEMA],
  bootstrap: [AppComponent],
})
export class AppModule {}
Vue

Install the package:

npm install @iblai/agent-ai

Use the component:

<template>
  <agent-ai
    mentorurl="https://mentor.iblai.app"
    authurl="https://auth.iblai.app"
    lmsurl="https://learn.iblai.app"
    tenant="your-tenant"
    mentor="mentor-id"
    isanonymous
    iscontextaware
    isadvanced
    theme="light"
  ></agent-ai>
</template>

<script>
export default {
  name: "App",
};
</script>

Context Sharing

This section describes the context sharing functionality implemented in the application, which allows for sharing HTML content between iframes and their parent windows.

Functions
  1. AgentAI.sendHTMLContentToHost(host: string, interval: number = 5000)

    This function is used when the application is running inside an iframe and needs to share the page content with the host. It creates a Web Worker to clean the HTML content before sending it to the host.

    • Parameters:

      • host: The target host to which the cleaned content will be sent.
      • interval: The time interval (in milliseconds) for checking and sending content (default is 5000 ms).
    • Usage:

      AgentAI.sendHTMLContentToHost("https://example.com", 5000);
  2. AgentAI.sendHTMLContentToIframe(iframeId: string, iframeHost: string, interval = 5000)

    This function allows the application to send HTML content to a specified iframe. It also utilizes a Web Worker for cleaning the content before sending it.

    • Parameters:

      • iframeId: The ID of the iframe element to which the content will be sent.
      • iframeHost: The host of the iframe.
      • interval: The time interval (in milliseconds) for checking and sending content (default is 5000 ms).
    • Usage:

      AgentAI.sendHTMLContentToIframe(
        "myIframe",
        "https://iframe-host.com",
        5000
      );
  3. AgentAI.proxyContextPostMessage(targetIdOrHost: string, host?: string)

    This function proxies post messages of type context either to an iframe or to the parent window, depending on the context in which it is called. It listens for incoming messages and forwards them accordingly.

    • Parameters:

      • targetIdOrHost: The ID of the target iframe or the host to which the message should be sent.
      • host: An optional parameter specifying the host to which the message will be sent.
    • Usage:

      AgentAI.proxyContextPostMessage("myIframe", "https://iframe-host.com");

Keywords