Post by Melissa Milloway

Here's how I use Claude Code + Claude Design to build PDFs that are on-brand, built for accessibility, and take seconds to regenerate. ➡️ It starts with a design and copy brief and a design system. I tell Claude what content the PDF needs to cover, who it's for, and what it should do. Then I sync my design system from Claude Design to Claude Code to implement it. One prompt pulls the design file in directly, fetches the design system file, reads the readme, implements the relevant parts and then Claude does the translation from design to code. ➡️ Because the layout is built in HTML, I can use Claude's frontend design skill alongside the design system from Claude Design. That means the PDF comes out with the right fonts, colors, and spacing, visual hierarchy, and a look that matches the rest of our materials. ➡️ One tool handles the conversion from HTML to PDF, and accessibility is built into how it runs. Some tools that turn web content into PDFs produce a flat file that screen readers can't navigate it. Claude Code uses a tool called WeasyPrint, a Python library that builds the underlying navigation structure a screen reader needs to move through a document. ➡️ Then I run every PDF through Adobe Acrobat's accessibility checker. I pull the full report and share it with Claude Code to fix what failed. But some things still need human review because automated checks don't catch everything. ➡️ Every accessibility failure I log in Claude Code. Like, a missing table header labels. And each one goes into a, skill MD document, I reuse so the next PDF automatically gets the accessibility rules applied. Have you tried this method to build well designed and more accessible PDFs? What else have you tried and how did it work? #eLearning #InstructionalDesign #LearningandDevelopment