Uprise SOP · Delivery

Website Framework Menu — which stack for which job

They are NOT one competing framework. They're two different layers, and they slot in differently:

source: notes/website-framework-menu.mdv1 — written the night the two Versorix docs got their first live test (ACOST redesign)created 2026-07-11

The verdict on the two Versorix docs (Rio asked 7/11)

They are NOT one competing framework. They're two different layers, and they slot in differently:

  1. design-copy-rulebook = the design brain. Adopt it for EVERY build, including previews, starting now. It is stack-agnostic (works on static HTML, Next.js, anything). Proof it earns its spot: on the ACOST redesign v1 it would have blocked exactly what Rio caught by eye ("looks like the Costa Co site"). Its anti-clone gate bans near-black + single bright accent + centered hero + grotesque-only type as the default AI look. The divergence engine + build ledger stop every site from rhyming.

  2. website-build-sop = the production pipeline (Claude Design → compiled static React → GitHub → Cloudflare Pages + SEO/AEO moat + tracking). It IS better than the current ad-hoc production approach for PAID client sites, because of what it adds after design: hidden per-page body-SEO (67→100 AEO score), llms.txt for AI crawlers, full schema, tracking wired as re-runnable scripts, phase gates so nothing ships half-verified. It is OVERKILL for previews and pitch demos. Cherry-pick its Phases 2-9 when there's a paying client and a real domain.

Bottom line: rulebook always, SOP when it's a production build for money.

The menu (pick per job, not per preference)

RIO'S CALL 2026-07-11: the SEO/AEO base layer is baseline on EVERY option below, previews included. Base layer = per-page meta + OG/Twitter tags + canonical, Schema.org JSON-LD, robots.txt, sitemap.xml, llms.txt, and the hidden body-SEO block (SOP §8.5). ~30-45 min per build. Only the ACCOUNT-side stack stays production-only (GA4/Pixel/Clarity wiring, Search Console, GBP) because it needs the client's accounts + a live domain.

# Framework When Time Output you get
1 Preview build — static HTML + CSS + vanilla JS, reference-driven, + SEO/AEO base layer Pitching / pre-close demos, ad-gap pages, teardown follow-ups 2-5 hrs Local or *.pages.dev link to text a prospect. Premium look AND the moat already baked in, so a closed deal graduates to production without rework
2 Production SOP stack — full website-build-sop Phases 0-9 Signed client, real domain, site that must rank + convert 8-16 hrs first time, 2-4 after Production site: 100/100 AEO + account-side tracking (GA4/Pixel/Clarity, GSC, GBP), phase-gated deploy on CF Pages
3 Next.js static export + SEO/AEO base layer Content-heavy sites: blog, many pages, CMS-fed (Dylan-style) varies Same CF Pages hosting, component reuse + MDX. Use only when page count / content ops demand it
4 Motion signature layer — GSAP ScrollTrigger + scroll-scrub video (SOP §4.11 + rulebook §8.4, Higgsfield for generated clips) Premium-positioning clients (film studios, luxury, launches). Layers ON TOP of #1 or #2, not a separate stack +2-4 hrs The "how is this not impossible" feel: ambient video hero, pinned scroll-scrub sections, scrubbed reveals. First live use: ACOST (hero ambient + statement scrub from Higgsfield clips of the client's own assets)

Hard rules carried from tonight's ACOST lessons (2026-07-11)

What still needs deciding (Rio)