Backend KOL_FEEDS trimmed from 29 to 25 (dead feeds removed).
Sync all hardcoded count mentions:
- layout.tsx JSON-LD, page.tsx (metric + comparison + copy)
- kol/page.tsx, KolPageClient.tsx ("and 26 more" → "and 22 more")
- glossary/page.tsx, opengraph-image.tsx
- public/llms.txt, llms-full.txt
- drop removed KOLs (Dragonfly Capital, Nic Carter) from named lists
Bundles other in-flight frontend work already in the working tree.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Trump:
- PageHint: "move price" → "move the market"
- Filter hints: "verdict is long/short BTC" → "scored as a long/short signal"
Archive:
- PageHint: "Historical fires from old scanner experiments" →
"Signals from retired scanner experiments … Read-only"
Case Studies:
- heroSubtitle: "worth studying" → "— documented."
- intro: remove redundant "is here to" phrasing
- footer: "signal stack" → "each signal engine"; "return promises" → "return forecasts"
Glossary:
- heroSubtitle: "The precise meaning of every key entity and term on the platform"
→ "Every term on the platform, defined precisely."
- intro: remove internal LLM-indexing rationale; rewrite for users
Methodology:
- heroSubtitle: "from raw input to trading decision"
→ "From raw data to trade decision — how each engine actually works."
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
GEO (AI answer engines) — make the signal feeds and case studies machine-citeable:
- Dataset JSON-LD on /trump, /kol, /trades (creator=Endorphin, publisher→#org,
isAccessibleForFree, temporalCoverage). These are the "data/history/track
record" entities Perplexity/Gemini/ChatGPT cite.
- case-studies: ItemList of 5 Article nodes, each with headline/description/
articleBody/about + citation→evidence URL + author=Endorphin. ISO dates
derived where parseable, omitted for vague labels.
- public/llms-full.txt: single-doc full reference (methodology + glossary +
case studies w/ sources inlined) for one-fetch AI ingestion; linked from llms.txt.
SEO:
- Breadcrumbs component (components/seo/Breadcrumbs.tsx) → BreadcrumbList JSON-LD
on the 8 indexable content pages (trump/kol/macro/trades/analytics/
methodology/glossary/case-studies).
- /archive split into server page + ArchivePageClient; server page sets
robots noindex (legacy/test data — thin-content risk).
- openGraph added to /trades and /analytics (previously meta+canonical only).
Verified: tsc 0 errors, next build passes, runtime curl confirms all JSON-LD
types render and llms-full.txt serves 200. No trading/API/component-logic changes.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
k