Work / Side Projects / James Website

Side Project — James Website

Rebuilt from a Notion page into a hand-coded site

jamespraise.xyz used to redirect to a Notion page through a Cloudflare Worker. I rebuilt it in code, the same way I rebuilt Marketing In Action's.

jamespraise.xyz
James Website screenshot
25pages built from scratch
11client case studies hosted natively instead of linking out
3competitor sites researched before deciding to rebuild

Live site ↗ · GitHub repo ↗

What It Is

jamespraise.xyz is my own site, this one. Hand coded in HTML, CSS, and plain JavaScript, the same way I rebuilt Marketing In Action's site. It replaced a Notion page that had been standing in as my site for years, redirected through a Cloudflare Worker.

Why I Built It

Notion let me ship a site without knowing how to code, which made sense back when I couldn't build one myself. Now that I run Claude Code every day, Notion's design ceiling became the actual problem, especially next to other operators in my own market who all have hand-coded sites of their own.

Before I touched a line of code, I did a proper teardown of a few peers in my own market. Each one does something a page builder can't: live animations, custom illustrated components, an AI chat widget answering questions for visitors, case studies with specific numbers attached. That teardown is what made rebuilding non-negotiable, not just a nice-to-have.

What It Does

The site runs my whole portfolio: home, work (client case studies, self-built side projects, and one-off pitch pages), about, speaking, and writing. Every client case study lives natively on the site now instead of linking out to a Notion database, the same fix I made for MIA's own case studies first. Every page shares one design system, built and refined the same way I refine anything else I ship.

What It Taught Me

Rebuilding my own site was the clearest proof yet that the skill I built doing this for MIA actually transfers. I didn't take a course or hire anyone. I ran the same process twice: research the competitors first, build in code with Claude Code, then keep shipping in small changes instead of one big rewrite.

It also taught me the actual limit of a tool like Notion. It's not speed, Notion was genuinely faster to start with. It's the ceiling on how far you can push design and interactivity before the tool itself is what's holding you back.