AI Extensions Framework Core Concepts

Contents

AI Extensions Framework Core Concepts#

The framework relies on some core concepts when configuring and using the plugin.

Core Concepts#

The plugin uses three main configuration concepts:

Provider

Handles authentication and model routing. Defines which AI service to use and how to connect to it. Set in Using Tutor Configuration (Recommended).

An example in Tutor’s config.yaml file:

AI_EXTENSIONS:
  provider:
    API_KEY: "sk-proj-your-api-key"
    MODEL: "provider/your-model"
Profile

Defines the what - what the AI will be instructed to do and which information it will have access to.

Example usage in Django Admin, where a profile is defined with a descriptive identifier (Slug) and an AI template to read from (Base filepath):

Profile configuration view showing base template and effective configuration
Scope

Defines the where - the context in which an AI workflow will be visible and usable (LMS/CMS, specific course, location, and frontend widget).

A scope has three matching dimensions:

  • UI Slot Selector ID: The id of the frontend widget that should render the workflow (see Using the UI Slot Selector ID).

  • Course ID: Optionally limits the scope to a specific course.

  • Location regex: Optionally limits the scope to specific units within a course.

Example usage in Django Admin, where a scope is defined with a widget ID, and optionally for a specific course or specific unit.

Create new scope interface

See also

Configuration Guide

Usage Guide

Setting Up a Local MCP Example Server

Model Context Protocol (MCP) Integration

Contents