Constructor Prism: September 8, 2026

Overview

In this release, we added the following feature:

• Display width controls for course blocks

We also fixed issues affecting media blocks, AI course generation, the course editor, the Content Library, course settings, enrollments, and scheduling.

What's new

Display width controls for course blocks

Authors can now select a block's display width from three icons in the course editor toolbar: text-column width, a wider content width, or full width.

The setting is available for images, image galleries, tables, code blocks, video, and Raw HTML blocks in media blocks, as well as video activities. It is applied in the course preview and learner view. For embedded Vimeo and YouTube videos, the setting resizes the video itself.

Display widths are relative to the text column, so they adjust to the available space on different screen sizes. Small images and videos remain at their original size instead of being stretched.

New blocks use the middle width setting by default. Existing blocks keep their current appearance. Blocks inside Toggles follow the width of the Toggle and do not have their own width control.

Fixed issues

We have addressed the following issues:

Media blocks and interactive video

• In-video quizzes now open from their timecodes, cannot be skipped through bookmarks or seeking, and show feedback correctly.
• Video bookmark labels, image galleries, contact cards, media-block captions, and long media titles now display correctly in previews and learner views.
• Images in course content now load more reliably, and misleading publishing-failure messages no longer appear after supported image or title updates.

AI course generation and publishing

• Generation progress now identifies the current step, and the interface resets correctly after a generation is cancelled.
• The agent now uses the current course structure after authors rename activities, modules, or lessons and update the brief.
• Video generation continues after an author leaves the course page, and generated videos display correctly after a course is published to Learn.
• Video player controls, publishing errors, draft switching, and SCORM package uploads now work more reliably.

Course editor and Content Library

• Content reviewers no longer enter an editing experience they cannot save, and unavailable VR lab editing and creation actions are no longer offered.
• Selecting, duplicating, moving, adding, and opening Content Library items now work more reliably.
• Content Library labels, loading indicators, image descriptions, dialogs, links, and tooltips now display more clearly and consistently.

Course settings, enrollments, and scheduling

• Course status, content-flow search and editing, expanded-tree state, score formatting, and editable-field feedback now update correctly.
• Enrollment settings, scheduling controls, certificate previews, and empty-course messaging have been corrected.

Known issues

The AI authoring limitations described in the August 11, 2026 release notes still apply. The following issues remain open:

• AI course generation currently supports PDF files and web links only. The file picker can offer document formats the agent does not yet read.
• Unsupported source files can be discarded without an error message.
• A course published to Learn can show incorrect passing-score and score-weight values.
• A course can be published even if a quiz in that course has not yet been generated.
• After an author edits the body text of a generated activity, the agent's answers do not yet reflect the edit.
• The agent can generate videos of at least two minutes even when a shorter clip is requested.