What's New

When rendering an avatar video, authors can now select either the Avatar III or Avatar IV engine each time — giving direct control over rendering quality and style.

Conversations can now start with the learner speaking first and a ringback tone before the AI answers — a more realistic setup for call-center training scenarios.

Enterprise admins can now see a last login column in the People list, with an exact timestamp visible on hover — making it easier to spot inactive accounts.

Make changes to your video instantly: edit text, images, and elements in place, add new scenes, insert shapes, and reorder overlapping elements without waiting on AI.

Mindsmith's interface is now available in Brazilian Portuguese (pt-BR), separate from European Portuguese, for authors and learners who prefer it.

A new public page lets IT teams check which Mindsmith domains their network can reach and copy a ready-to-share list of any that are blocked.

Design your organization's completion certificate with a custom template, background image, and control over which elements — including the verification link — appear.

Expert mode now runs on Claude Opus 4.8, Anthropic's most capable model, with no action required from your organization.

The narration voice picker now lets you filter by accent — US, UK, Australian English, and more — so finding the right voice takes far fewer listens.

A new captions menu in Mindsmith Video V2 lets editors add, download, or remove captions, and download the video with a matching caption file.

Edit a scene's image prompt before regenerating, and preview all scenes as one continuous playback — both new in Mindsmith Video V2.

AI tools connected via MCP can now find lessons by course or folder name and create AI practice conversation lessons directly.

Images generated by AI in lessons complete roughly twice as fast at the same credit cost.

Editable video can render with a new, higher-quality AI avatar engine, and render status now shows on the tile — with failed re-renders keeping the last working version.

Rename, recolor, or delete content tags from Settings, with a usage count shown before you delete one.

Importing Articulate Storyline course archives is now available to all authors — no special account setting required.

The AI agent can now edit stock images and author-uploaded images in a lesson, not just the ones it generated itself.

SCORM 2004 exports now include per-question interaction data — the question text, the learner's response, and whether it was correct — so LMSs like Cornerstone and SCORM Cloud can report on individual answers, not just the overall score.

The conversation create page now offers ready-made starting templates alongside eight built-in B2B sales skills — discovery, objection handling, deal qualification, and more — for AI-powered role-plays.

Org admins can see total AI conversation sessions, minutes used, and estimated cost in settings — and learners see a clear message when the org limit is reached.