The animation modes are AI Motion Graphics, Avatar video, Screencast, and Screencast + Avatar. Inputs can be an idea, script, URL, document, image, audio clip, or screen recording. For product walkthroughs, the tool detects clicks, smooths the cursor, zooms to the action, and trims dead air. On the avatar side, four sources are live: a built-in library, generated on-brand protagonists, approved faces in a workspace library, and uploaded face images or short clips with lip sync. Avatar-first tools such as Synthesia and HeyGen start from a script, and ngram starts from existing material like docs and recordings. One review notes that Synthesia or HeyGen still have deeper avatar libraries, while another says ngram is a general-purpose video creation tool and is not tuned for demos specifically. A third-party review also calls it worth the investment for SaaS teams that regularly ship demos and feature videos. For founders building AI products, this is a case of video generation becoming a workflow layer rather than a model feature. The competitive ground is shifting from avatar quality to how well a tool turns messy source material, such as raw recordings and docs, into usable output. Teams selling into product marketing should expect buyers to compare tools on editing and revision speed, not on the avatar alone.