From bd76ff5b13d88f50327b8c0d2f1f9194adda5a09 Mon Sep 17 00:00:00 2001
From: Flotapponnier <160007691+Flotapponnier@users.noreply.github.com>
Date: Sun, 26 Jul 2026 07:02:09 +0200
Subject: [PATCH] seo P2 batch: CSP+twitter:site+grammar+hyperliquid
 Dataset+feed.json (#1490)
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

Five P2 fixes from the SEO deep audit (2026-07-26):

1. CSP: allowlist analytics.ahrefs.com in script-src + connect-src.
   Ahrefs script was silently violating CSP on every visitor.
2. products/[slug]: add twitter:site (was only card+title+description
   before, every other route type ships it).
3. chains/[slug]: pluralize 'N measurement' vs 'N measurements'
   ('1 measurements' on Polkadot hub reads broken).
4. hyperliquid/[slug]: add Dataset JSON-LD (only had BreadcrumbList,
   GSC treats content-heavy pages with only breadcrumbs as 'no
   structured data'). Dataset carries revenue+volume variableMeasured
   pointing at /api/stat/hyperliquid-frontends.
5. Add /feed.json (JSON Feed 1.1) as a companion to /rss.xml. Modern
   readers + AI agent tooling (Perplexity crawler, LangChain document
   loaders, MCP clients) prefer JSON over XML. <link rel=alternate>
   added in layout for auto-discovery.

Skipped items from the audit:
- SITE.github Flotapponnier: false positive, ChainBench URL works.
- robots.txt User-Agent casing: both forms are spec-compliant.
- /ai.txt: no formal spec exists; llms.txt already covers.
- JSON-LD body→head migration: refactor cost > payoff.
- MCP directory registration: external, requires Anthropic account.

Co-authored-by: Florent Tapponnier <contact@mobula.io>
---
 next.config.ts                      | 11 +++-
 src/app/chains/[slug]/page.tsx      |  2 +-
 src/app/feed.json/route.ts          | 92 +++++++++++++++++++++++++++++
 src/app/hyperliquid/[slug]/page.tsx | 50 +++++++++++++++-
 src/app/layout.tsx                  |  9 +++
 src/app/products/[slug]/page.tsx    |  2 +-
 6 files changed, 160 insertions(+), 6 deletions(-)
 create mode 100644 src/app/feed.json/route.ts

diff --git a/next.config.ts b/next.config.ts
index a68ed197..0b0f6e15 100644
--- a/next.config.ts
+++ b/next.config.ts
@@ -12,19 +12,24 @@ const RELAY_WS = "wss://stream.openchainbench.com";
 // modal's <video> tag can play the result, and in connect-src so the
 // proxy's success URL can be opportunistically pre-fetched.
 const VIDEO_RENDERER_ORIGIN = "https://video.openchainbench.com";
+// Ahrefs Web Analytics. Loaded from analytics.ahrefs.com and beacons
+// back to the same origin — needs both script-src and connect-src
+// allowlist entries. Kept as a distinct constant so a future analytics
+// swap is a one-line change.
+const AHREFS_ORIGIN = "https://analytics.ahrefs.com";
 const IS_DEV = process.env.NODE_ENV !== "production";
 // React dev mode needs 'unsafe-eval' for fast refresh / call stack reconstruction.
 // In prod we keep the lock-tight policy.
 const SCRIPT_SRC = IS_DEV
-  ? "script-src 'self' 'unsafe-inline' 'unsafe-eval'"
-  : "script-src 'self' 'unsafe-inline'";
+  ? `script-src 'self' 'unsafe-inline' 'unsafe-eval' ${AHREFS_ORIGIN}`
+  : `script-src 'self' 'unsafe-inline' ${AHREFS_ORIGIN}`;
 const CSP = [
   "default-src 'self'",
   SCRIPT_SRC,
   "style-src 'self' 'unsafe-inline'",
   "img-src 'self' data: blob:",
   "font-src 'self' data:",
-  `connect-src 'self' ${RELAY_WS} ${VIDEO_RENDERER_ORIGIN}`,
+  `connect-src 'self' ${RELAY_WS} ${VIDEO_RENDERER_ORIGIN} ${AHREFS_ORIGIN}`,
   `media-src 'self' ${VIDEO_RENDERER_ORIGIN}`,
   "frame-ancestors 'none'",
   "base-uri 'self'",
diff --git a/src/app/chains/[slug]/page.tsx b/src/app/chains/[slug]/page.tsx
index 88e1e7c5..6c8db84c 100644
--- a/src/app/chains/[slug]/page.tsx
+++ b/src/app/chains/[slug]/page.tsx
@@ -50,7 +50,7 @@ export async function generateMetadata({
   const url = `${SITE.url}/chains/${slug}`;
   const title = `${chain.label} live benchmarks: finality, fees, RPC`;
   const description = capDescription(
-    `${benches.length} live OpenChainBench measurements covering ${chain.label}. ${chain.description}`,
+    `${benches.length} live OpenChainBench measurement${benches.length === 1 ? "" : "s"} covering ${chain.label}. ${chain.description}`,
     158,
   );
   return {
diff --git a/src/app/feed.json/route.ts b/src/app/feed.json/route.ts
new file mode 100644
index 00000000..7ec5c46f
--- /dev/null
+++ b/src/app/feed.json/route.ts
@@ -0,0 +1,92 @@
+/**
+ * JSON Feed 1.1 (https://jsonfeed.org/version/1.1) mirror of /rss.xml.
+ *
+ * Modern feed readers + several AI agent tools (Perplexity's crawler,
+ * LangChain document loaders, MCP clients) prefer JSON Feed over RSS
+ * because JSON is trivial to parse without an XML dependency. RSS
+ * stays as the canonical feed for legacy aggregators; feed.json ships
+ * the same entries as a companion so tooling that only speaks JSON
+ * also picks up new benches without a custom scraper.
+ *
+ * Same freshness contract as rss.xml: pubDate per item = bench first
+ * commit; feed-level `date_modified` = generation time so aggregators
+ * see the feed as fresh whenever a headline number changes. Cache 5 min
+ * edge TTL, same as RSS.
+ */
+
+import { NextResponse } from "next/server";
+import { AllBenchmarksDraftError, loadAllBenchmarks } from "@/lib/spec";
+import { getBenchCreatedAt } from "@/lib/seo/bench-dates";
+import { headlineSentence } from "@/lib/citation";
+import { SITE } from "@/data/site";
+import type { Benchmark } from "@/types/benchmark";
+
+export const revalidate = 300;
+
+function itemContentText(b: Benchmark): string {
+  const sentence = headlineSentence(b);
+  if (sentence) return `${sentence} ${b.subtitle}`.trim();
+  return b.subtitle;
+}
+
+export async function GET() {
+  let all;
+  try {
+    all = await loadAllBenchmarks();
+  } catch (err) {
+    if (err instanceof AllBenchmarksDraftError) {
+      return NextResponse.json(
+        { error: "benchmarks_unavailable" },
+        {
+          status: 503,
+          headers: {
+            "cache-control": "no-store",
+            "retry-after": "60",
+          },
+        },
+      );
+    }
+    throw err;
+  }
+  const live = all.filter((b) => b.editorialStatus === "live");
+
+  const items = live
+    .map((b) => ({ bench: b, pubDate: getBenchCreatedAt(b.slug) }))
+    .sort((a, c) => c.pubDate.getTime() - a.pubDate.getTime())
+    .map(({ bench: b, pubDate }) => {
+      const url = `${SITE.url}/benchmarks/${b.slug}`;
+      return {
+        id: url,
+        url,
+        title: b.title,
+        content_text: itemContentText(b),
+        summary: b.subtitle,
+        date_published: pubDate.toISOString(),
+        ...(b.lastRunAt ? { date_modified: b.lastRunAt } : {}),
+        tags: [b.category],
+        authors: [{ name: "OpenChainBench", url: SITE.url }],
+      };
+    });
+
+  const feed = {
+    version: "https://jsonfeed.org/version/1.1",
+    title: "OpenChainBench benchmark releases",
+    home_page_url: SITE.url,
+    feed_url: `${SITE.url}/feed.json`,
+    description:
+      "Live measurements for crypto infrastructure: RPCs, oracles, aggregators, bridges, prediction markets. One entry per public benchmark.",
+    language: "en",
+    icon: `${SITE.url}/logo.png`,
+    favicon: `${SITE.url}/icon`,
+    authors: [{ name: "OpenChainBench", url: SITE.url }],
+    date_modified: new Date().toISOString(),
+    items,
+  };
+
+  return NextResponse.json(feed, {
+    headers: {
+      "content-type": "application/feed+json; charset=utf-8",
+      "cache-control": "public, s-maxage=300, stale-while-revalidate=3600",
+    },
+  });
+}
diff --git a/src/app/hyperliquid/[slug]/page.tsx b/src/app/hyperliquid/[slug]/page.tsx
index 231318dc..b41ec784 100644
--- a/src/app/hyperliquid/[slug]/page.tsx
+++ b/src/app/hyperliquid/[slug]/page.tsx
@@ -121,6 +121,7 @@ export default async function HlFrontendPage({
 
   const peers = pickPeers(history.frontends, frontend, 5);
 
+  const pageUrl = `https://openchainbench.com/hyperliquid/${slug}`;
   const breadcrumbLd = {
     "@context": "https://schema.org",
     "@type": "BreadcrumbList",
@@ -141,7 +142,49 @@ export default async function HlFrontendPage({
         "@type": "ListItem",
         position: 3,
         name: frontend.name,
-        item: `https://openchainbench.com/hyperliquid/${slug}`,
+        item: pageUrl,
+      },
+    ],
+  };
+
+  // Content-shape schema. Previously only BreadcrumbList was emitted,
+  // which GSC treats as "no structured data for the primary content"
+  // on data-heavy pages (frontend revenue + volume timeseries). Add a
+  // Dataset node so Google Dataset Search + LLM crawlers index the
+  // 30-day revenue/volume figures as a citable measurement rather than
+  // treating the page as generic prose.
+  const datasetLd = {
+    "@context": "https://schema.org",
+    "@type": "Dataset",
+    "@id": `${pageUrl}#dataset`,
+    name: `${frontend.name} — HyperLiquid frontend revenue + volume`,
+    description: `Daily HyperLiquid frontend fees and volume for ${frontend.name}, sourced from the OpenChainBench hyperliquid-frontends benchmark. First measured ${firstDay}.`,
+    url: pageUrl,
+    identifier: slug,
+    creator: {
+      "@type": "Organization",
+      "@id": "https://openchainbench.com/#org",
+      name: "OpenChainBench",
+      url: "https://openchainbench.com",
+    },
+    publisher: {
+      "@type": "Organization",
+      "@id": "https://openchainbench.com/#org",
+      name: "OpenChainBench",
+      url: "https://openchainbench.com",
+    },
+    isAccessibleForFree: true,
+    license: "https://creativecommons.org/licenses/by/4.0/",
+    variableMeasured: [
+      "HyperLiquid frontend fees (30d)",
+      "HyperLiquid frontend volume (30d)",
+    ],
+    temporalCoverage: `${new Date(firstDayMs).toISOString().slice(0, 10)}/..`,
+    distribution: [
+      {
+        "@type": "DataDownload",
+        encodingFormat: "application/json",
+        contentUrl: "https://openchainbench.com/api/stat/hyperliquid-frontends",
       },
     ],
   };
@@ -153,6 +196,11 @@ export default async function HlFrontendPage({
         // biome-ignore lint/security/noDangerouslySetInnerHtml: serialized via safeJsonLd
         dangerouslySetInnerHTML={{ __html: safeJsonLd(breadcrumbLd) }}
       />
+      <script
+        type="application/ld+json"
+        // biome-ignore lint/security/noDangerouslySetInnerHtml: serialized via safeJsonLd
+        dangerouslySetInnerHTML={{ __html: safeJsonLd(datasetLd) }}
+      />
 
       <Breadcrumb
         items={[
diff --git a/src/app/layout.tsx b/src/app/layout.tsx
index 14d36da7..450ff7bd 100644
--- a/src/app/layout.tsx
+++ b/src/app/layout.tsx
@@ -201,6 +201,15 @@ export default async function RootLayout({
           title="OpenChainBench — new benchmarks"
           href="/rss.xml"
         />
+        {/* JSON Feed 1.1 companion. Modern feed readers + AI agent
+            tooling (Perplexity, LangChain document loaders, MCP
+            clients) parse JSON without an XML dependency. */}
+        <link
+          rel="alternate"
+          type="application/feed+json"
+          title="OpenChainBench — new benchmarks (JSON Feed)"
+          href="/feed.json"
+        />
         <script
           dangerouslySetInnerHTML={{
             __html: `(function(){try{var t=localStorage.getItem('ocb-theme');var d=t==='dark'||(!t&&window.matchMedia('(prefers-color-scheme: dark)').matches);if(d)document.documentElement.classList.add('dark');}catch(e){}})();`,
diff --git a/src/app/products/[slug]/page.tsx b/src/app/products/[slug]/page.tsx
index 55dc11f6..9ba71f1f 100644
--- a/src/app/products/[slug]/page.tsx
+++ b/src/app/products/[slug]/page.tsx
@@ -144,7 +144,7 @@ export async function generateMetadata({
     description,
     alternates: { canonical: canonicalUrl },
     openGraph: { title, description, type: "profile", url: canonicalUrl },
-    twitter: { card: "summary_large_image", title, description },
+    twitter: { card: "summary_large_image", site: SITE.twitter, title, description },
   };
 }