This section describes the Settings panel available inside the course, focusing on AI-powered features and Notifications management. These settings allow course creators to automate lecture summaries, improve lecture naming, and control system notifications related to course activity.1
Access Course SettingsNavigate to the Settings section of the course.
The Settings panel opens, displaying AI and Notifications options.
2
AI SettingsEnable this option to automatically generate a clear and concise summary of each lecture based on the AI transcription.
The generated summary is also saved as a PDF and attached to the lesson.
An optional custom prompt can be added to refine how the AI generates the summary (up to 500 characters).
Enable this option to automatically rename lectures using AI.
The new lecture title reflects the actual lecture content derived from the transcription.
3
Notification SettingsRecording and AI Summary GeneratedEnable to receive notifications when a lecture recording and its AI-generated summary are ready.
Enable to be notified when a lecture is rescheduled.
Enable to receive alerts when a lecture is suspended.
How the AI Summary Prompt Works#
When the system generates an AI summary, it follows a set of instructions called a prompt.
This prompt defines what kind of summary the AI writes and how it is structured.Default Prompt (Standard Behavior)Used automatically when no custom prompt is provided.
Create a clear lesson title
Write a short introduction
Break the lecture into main themes (macro topics)
Explain each theme in well-written paragraphs (not bullet points)
Include links to reliable academic sources
End with a short conclusion
This ensures summaries are structured, academic, consistent, and easy to study from.
Used only when a custom prompt is added.
Replaces the default instructions entirely.
Shorter or longer summaries
Focus on practical takeaways
Only one prompt is used at a time.
The AI always reads the full lecture transcription.
The summary remains in the same language as the lecture.
The output stays structured, predictable, and machine-readable.
Modified at 2025-12-24 13:56:43
