You’re right—we can (and should) look deeper into the *technical* underpinnings of the Kayba site to extract lessons for our own build. Here’s a more granular breakdown of the competitor’s **design stack** and layout choices, plus how we can outshine them with our React/Tailwind stack.
---
## 🔬 What Kayba.ai is doing technically
* **Built with Framer:** The HTML source explicitly identifies Framer as the site generator—there’s a script loading `framer.com/edit/init.mjs` and a `<meta name="generator" content="Framer …>` tag. This indicates the entire site is published via Framer’s no‑code platform, which provides visual design tools and built‑in CMS.
* **Custom fonts & gradients:** They load multiple weights of **DM Sans** and **Inter** from Google Fonts. These fonts, paired with layered gradient backgrounds and a subtle star field, create their “midnight zone” aesthetic.
* **Section layouts:** The page is a long scroll with:
* A hero block set against a cosmic gradient (“The Context Layer Your AI Is Missing”).
* A “Why agentic systems fail” section with three cards explaining poor retrieval, context rot, and unstructured context.
* A features section introducing **Token Compaction**, **Context Orchestration**, and **Retrieval & Prioritization**.
* A diagr…