Anyone who has tried to make slides for a video, presentation, or course has run into the same bottleneck: the content exists in your head or in a script, but turning it into something visual takes a disproportionate amount of time compared to how long it took to write.
One creator demonstrated a specific workflow for solving this using Claude Code, and the method is straightforward enough to apply without any design background at all.
The starting point: one point per slide
Before any technical setup, the underlying principle matters more than the tool. A slide should carry one idea, not several. The instinct when creating slides is often to cram in everything related to a topic, producing a dense, text-heavy slide that is difficult to read at a glance and does little to support whoever is speaking over it.
The better approach treats each slide as a single visual anchor for a single point. If your script moves through five distinct ideas, that becomes five slides, not one overloaded slide trying to hold all five.
Setting up the workflow
The practical process starts by using Claude Code’s skill-creation feature, which allows you to teach Claude a repeatable task once and then reuse it going forward. In this case, the skill being built is specifically for generating slides from a script.
Once that skill exists, the workflow becomes simple from the user’s side: provide the script, and ask Claude to apply the “one idea per slide” rule, specifying that each slide should correspond to a single clear point rather than a block of related information.
Claude then generates the slide content and structure, which gets built out as an actual file — ready to be dropped into whatever editing or presentation tool you’re using, rather than existing only as a text description of what a slide should contain.
Why this works better than starting from a blank slide deck
The value here is less about automation for its own sake, and more about removing the specific bottleneck that makes slide creation slow: deciding how to visually break down written content. That decision-making step, repeated across every slide in a deck, is where most of the time goes. Having Claude apply a consistent rule — one point, one slide — removes a significant amount of that repeated decision-making.
This does not replace the judgement needed to review and refine the output. Slides generated this way are a strong starting point, not a finished product. Reviewing for accuracy, adjusting wording to match your actual speaking style, and checking that the visual breakdown genuinely matches how you want to present the material all remain necessary steps.
What changes is the starting position. Instead of facing a blank deck and building from nothing, you are refining something that already follows a sound structural principle, built in a fraction of the time manual slide creation would normally take.
Where this applies beyond video
Although the original example was built around YouTube video production, the same workflow applies to any situation requiring a script-to-slides conversion: client presentations, internal training material, course content, or conference talks. The underlying method — teach Claude the one-point-per-slide rule once, then apply it repeatedly — works regardless of the final destination for the slides.
If you want to go further and build genuinely reusable Skills for repeated tasks like this one, Claude PowerPrint covers Skills, Artifacts and the wider Claude toolset properly.
