Key Improvements

Expanded OpenAI API Integration

• Chat API – Standard conversational AI.
• Responses API – Includes:
    - Moodle API integration
    - File-based contextual search
• Assistant API – Includes:
    - Moodle API integration
    - File-based contextual search
• Azure (Self-hosted OpenAI) – Supports Azure-hosted OpenAI instances.

Main Features

• Moodle API Integration
• The AI can dynamically interact with Moodle APIs to fulfill user requests.
  Example:
  A user asks for their course grades — the AI calls the Moodle Grades API and provides a personalized response.
• File Search Support:
  Admins can upload files which are made available to the AI during chats. This enables context-aware responses based on the provided materials.
• Custom User Messages with Dynamic Placeholders: Each chat instance can define a custom initial message to send to the AI, supporting dynamic user/course-based placeholders.
  Example message template:
  "My name is {user.fullname} and my final course grade is {grade:Final Grade}."
  
  These placeholders are replaced with actual user and course data before being sent.
  As the AI is not aware of the current date/time, you can also add a placeholder for the current date/time, such as `{userdate}`.
  
  Available Placeholders:
  {user.fullname} - full name of the user
  {userdate} - current date/time in the user's timezone
  {grade:entryname} - get the grade from the gradebook with the name of "entryname". entryname can be any item from the gradebook, e.g., "Final Grade", "Midterm Exam", or any activity name
  
  
• Detailed API Logging (Optional):
  Enable fine-grained logging of all API requests and responses for in-depth debugging and diagnostics of backend interactions with the AI providers.

Planned Features

• Integration of additional AI providers beyond OpenAI.
• Recording attendance through Moodle's Attendance block
• Extend scope of placeholder
• check group membership

The development of the block was carried out by GTN in collaboration with the Laboratory for Advanced Technology in Education at the University of Zagreb Faculty of Organization and Informatics. The Laboratory contributed to the plugin’s conceptual design, pedagogical framework, and evaluation process, bringing extensive expertise in educational technology and its application within higher education contexts.

For more information, visit: https://www.foi.unizg.hr/en/about-us/departments/late