Summary
This tutorial from Jack Roberts shows how to use Claude to produce designer-quality motion graphics and short videos from plain-English descriptions. It opens with examples such as paper-world scenes, a growing city, Swiss kinetic typography and an interactive sky widget, all generated by Claude with no manual editing.
The central argument is that identical prompts produce identical, generic results, which is why so much AI-made video looks like “slop.” Roberts compares one-shot output with results produced using a reusable skill. The skill-guided versions add depth of field, interactivity and more deliberate visual styling. He walks through a repeatable process: pick a visual style, gather real design references, and give Claude a skill to follow by sharing its URL.
A second major section covers mixing generated motion graphics with real photos and video footage. Side-by-side comparisons show how adding a personal photo or live footage turns a generic animation into something that looks bespoke, and how viewers can switch between graphic and media at any point. Roberts also points to uses in pitch decks, product explainers and website widgets, and offers a free skill for viewers to follow along. The video suits creators, marketers and founders who want a repeatable system for producing motion content with Claude.
📺 Source: Jack Roberts · Published October 24, 2026
🏷️ Format: Tutorial Demo
