What's New

A single filmstrip spanning every page across the whole course stays visible while you edit, so you can move between pages and lessons in one session.

Authors can now add uploaded video, screen recordings, or stock clips as b-roll to AI video scenes, with in-editor trim and volume controls.

Image generation now reads the surrounding page text, adapts to how the image will be used, and draws on your organization's theme colors.

The agent asks about image style, how to use attached documents, and — for video — presenter type and visual style, before it starts building.

PDF exports of lessons and courses can now include reviewer comments, with options to show only unresolved comments or a single language.

New lesson prose avoids the usual AI writing tells by default, and a selectable humanizer pass rewrites an existing lesson's copy in one action.

Every Video V2 render now comes back with captions and a transcript, org-uploaded characters can present, and non-Latin scripts render correctly.

Paid teams can now choose ElevenLabs' higher-quality v3 narration voice from an account setting, in place of the default model.

The media library supports selecting multiple files at once to tag, download, or delete, with a new list view and larger previews.

Learners who reach a teamspace's AI conversation limit can now finish the lesson and course instead of being stuck on an incomplete tile.

The course-level AI assistant can now add narration, edit the course theme, and run diagnostics across all lessons without opening each one individually.

Dashboard content can now be dragged directly into folders and courses, and the move dialog lets you create new folders without leaving the dialog.

Videos that ran past the old ten-minute render ceiling — long lessons and anything with an avatar — now complete instead of failing at the end.

Glossary terms for Latin American Spanish and French Canadian are now applied correctly when translating, rather than falling back to shared base-language terms.

The shape tool adds circles, triangles, comment bubbles, and arrows, and the timeline now marks scene boundaries so playback starts from the scene you selected.

AI tools connected via MCP can now discover and browse courses by title or folder, and retrieve a course's full lesson outline.

Paste images straight into the chat, add a tile by typing its name, and follow long edits through a cleaner activity log.

The AI agent now reads the rest of your course before editing a lesson, and can see the images already placed inside it.

The AI editing agent can now update a lesson's narration script, change the narrator voice, and toggle narration on or off.

Fast and Expert modes are now Lite and Max, with a redesigned mode selector, Max mode available to Pro and Team plans, and Claude Opus 5 powering Max.