From a9a891cc3f20638f4da15d32dfa04c4f327e239c Mon Sep 17 00:00:00 2001
From: Prabal <prabal@availproject.org>
Date: Fri, 4 Sep 2026 16:44:28 +0400
Subject: [PATCH 1/3] =?UTF-8?q?v2.3.0=20=E2=80=94=20Fix=20agentic=20readab?=
 =?UTF-8?q?ility=20blockers=20(is-agentic=2064/100)?=
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

Scanned docs.availproject.org against is-agentic.com and fixed the
failing checks. No human-facing content or layout was altered; the three
new pages were added at the maintainer's request.

Server-rendered content (Essential — was failing)
Every page served ~0 chars of content in raw HTML. NexusContainer
returned a full-page <Skeleton> until a useEffect read localStorage, and
Web3Provider wraps the whole app, so the server emitted a skeleton for
the entire body. Existing canonical URLs and TechArticle/BreadcrumbList
JSON-LD never reached the HTML either.

NexusProvider builds its client inside useEffect([config]) with destroy()
cleanup, so it is SSR-safe and rebuilds correctly when the stored network
differs from the default. Dropping the gate is therefore safe.

  /                     35 ->  1584 chars
  /docs/da/get-started  54 ->  2541 chars
  /docs/nexus           40 ->   966 chars

Markdown content negotiation (Essential — was failing)
- middleware rewrote /docs/da/* to /api/markdown/DA/*, but content is
  canonicalized lowercase, so every Avail DA page 404'd on both the
  Accept: text/markdown and .md paths. Uppercase /api/markdown/DA/* 404'd
  as well. Both casings now resolve.
- Vary was emitted as a second header line, which CDNs and scanners
  commonly miss. Now one merged value, with Accept first.

Agent-friendly 404s (Essential — was partial)
404s return a markdown body naming the discovery indexes instead of an
opaque JSON error, keeping the 404 status.

Structured data and metadata (Recommended — were failing)
- Homepage: canonical URL plus an Organization/WebSite/SoftwareApplication
  @graph, with contactPoint and sameAs built only from links already in
  the repo.
- Docs pages: the 3-field Organization stub replaced with the full node.

Discoverability
- /about, /contact, /privacy added (previously 404). Privacy documents
  only verified behaviour, and discloses that feedback is filed to a
  public GitHub repo.
- sitemap.xml now includes the homepage and these pages; it previously
  listed docs pages only.
- llms.txt gains an Agent Endpoints section.

Tests updated: the markdown API and middleware suites asserted the
uppercase-DA behaviour that was breaking production. 100 vitest + 1 jest
pass; Biome error count unchanged from main (121).

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
---
 src/app/(home)/about/page.tsx             | 71 ++++++++++++++++
 src/app/(home)/contact/page.tsx           | 71 ++++++++++++++++
 src/app/(home)/page.tsx                   | 11 +++
 src/app/(home)/privacy/page.tsx           | 99 +++++++++++++++++++++++
 src/app/api/markdown/[...slug]/route.ts   |  9 ++-
 src/app/api/markdown/markdown-api.test.ts | 11 ++-
 src/app/api/markdown/route.ts             |  3 +-
 src/app/docs/[[...slug]]/page.tsx         | 11 ++-
 src/app/not-found.tsx                     | 18 ++++-
 src/app/sitemap.test.ts                   | 16 +++-
 src/app/sitemap.ts                        | 21 ++++-
 src/components/info/info-page.tsx         | 63 +++++++++++++++
 src/lib/agent-headers.ts                  | 12 ++-
 src/lib/agent-not-found.ts                | 49 +++++++++++
 src/lib/footer-links.ts                   |  3 +
 src/lib/llms.ts                           | 17 ++++
 src/lib/structured-data.ts                | 92 +++++++++++++++++++++
 src/middleware.test.ts                    | 27 +++++--
 src/middleware.ts                         | 41 ++++------
 src/providers/Web3Provider.tsx            | 12 +--
 20 files changed, 595 insertions(+), 62 deletions(-)
 create mode 100644 src/app/(home)/about/page.tsx
 create mode 100644 src/app/(home)/contact/page.tsx
 create mode 100644 src/app/(home)/privacy/page.tsx
 create mode 100644 src/components/info/info-page.tsx
 create mode 100644 src/lib/agent-not-found.ts
 create mode 100644 src/lib/structured-data.ts

diff --git a/src/app/(home)/about/page.tsx b/src/app/(home)/about/page.tsx
new file mode 100644
index 0000000..2173b31
--- /dev/null
+++ b/src/app/(home)/about/page.tsx
@@ -0,0 +1,71 @@
+import { InfoPage, InfoSection } from "@/components/info/info-page";
+import { SITE_URL } from "@/lib/structured-data";
+
+export const metadata = {
+  title: "About",
+  description:
+    "About the Avail documentation site — what Avail DA and Avail Nexus are, who maintains these docs, and how the content is organised.",
+  alternates: { canonical: `${SITE_URL}/about` },
+};
+
+export default function AboutPage() {
+  return (
+    <InfoPage
+      title="About these docs"
+      lede="docs.availproject.org is the official documentation for Avail, published and maintained by the Avail Project."
+    >
+      <InfoSection heading="What Avail is">
+        <p>
+          Avail is a modular blockchain stack. It is split into two products,
+          and this site documents both of them.
+        </p>
+        <p>
+          <strong>Avail DA</strong> is a data availability layer. Rollups and
+          other chains post their transaction data to Avail DA and can prove
+          that the data was published, without every participant having to
+          download all of it. The docs cover submitting and reading data,
+          running full nodes and validators, light clients, network endpoints
+          and chain specs, and the node API across TypeScript, Rust and Go.
+        </p>
+        <p>
+          <strong>Avail Nexus</strong> is a cross-chain unification layer. It
+          gives applications a single balance and a single transaction flow
+          across many chains, through an SDK or through drop-in UI widgets. The
+          docs cover the SDK reference, bridging, swaps, supported chains and
+          tokens, and the widget components.
+        </p>
+      </InfoSection>
+
+      <InfoSection heading="How this site is organised">
+        <p>
+          Content is split by product: everything under <code>/docs/da</code> is
+          Avail DA, and everything under <code>/docs/nexus</code> is Avail
+          Nexus. Each product has its own Get Started path, conceptual material,
+          and API reference.
+        </p>
+      </InfoSection>
+
+      <InfoSection heading="For AI agents and scripts">
+        <p>
+          These docs are published in machine-readable form as well as HTML.{" "}
+          <a href="/llms.txt">/llms.txt</a> is an intent-based index of the
+          whole site, with per-product dumps at{" "}
+          <a href="/llms-da.txt">/llms-da.txt</a> and{" "}
+          <a href="/llms-nexus.txt">/llms-nexus.txt</a>. Any documentation URL
+          returns clean markdown if you append <code>.md</code> to it or send an{" "}
+          <code>Accept: text/markdown</code> header, and structured reference
+          data is available at{" "}
+          <a href="/api/reference.json">/api/reference.json</a>.
+        </p>
+      </InfoSection>
+
+      <InfoSection heading="Corrections and contributions">
+        <p>
+          Every page has a feedback control, and the docs are open source. If
+          something here is wrong or missing, please tell us — see{" "}
+          <a href="/contact">Contact</a> for the ways to reach the team.
+        </p>
+      </InfoSection>
+    </InfoPage>
+  );
+}
diff --git a/src/app/(home)/contact/page.tsx b/src/app/(home)/contact/page.tsx
new file mode 100644
index 0000000..873ea23
--- /dev/null
+++ b/src/app/(home)/contact/page.tsx
@@ -0,0 +1,71 @@
+import { InfoPage, InfoSection } from "@/components/info/info-page";
+import { SITE_URL } from "@/lib/structured-data";
+
+export const metadata = {
+  title: "Contact",
+  description:
+    "How to reach the Avail team — developer support on Discord, documentation issues on GitHub, security disclosures, and business enquiries.",
+  alternates: { canonical: `${SITE_URL}/contact` },
+};
+
+export default function ContactPage() {
+  return (
+    <InfoPage
+      title="Contact"
+      lede="How to reach the Avail team, depending on what you need."
+    >
+      <InfoSection heading="Developer support">
+        <p>
+          For build questions, integration help, and anything about running
+          nodes, the fastest route is the community. Ask in the{" "}
+          <a href="https://discord.com/invite/AvailProject">Avail Discord</a>,
+          where the engineering team is active, or in the{" "}
+          <a href="https://t.me/AvailCommunity">Avail Telegram community</a>.
+        </p>
+      </InfoSection>
+
+      <InfoSection heading="Documentation issues">
+        <p>
+          If a page is wrong, out of date, or missing, use the feedback control
+          at the bottom of that page — it goes straight to the docs backlog. You
+          can also open an issue in the{" "}
+          <a href="https://github.com/availproject">
+            Avail GitHub organisation
+          </a>{" "}
+          against the relevant repository.
+        </p>
+      </InfoSection>
+
+      <InfoSection heading="Security disclosures">
+        <p>
+          Please do not report vulnerabilities in public channels. Email{" "}
+          <a href="mailto:security@availproject.org">
+            security@availproject.org
+          </a>
+          . Scope, rewards, and the disclosure process are described in the{" "}
+          <a href="/docs/da/bug-bounty">bug bounty documentation</a>.
+        </p>
+      </InfoSection>
+
+      <InfoSection heading="Business and partnerships">
+        <p>
+          For partnership, integration, and commercial enquiries — including
+          access to Turbo DA — email{" "}
+          <a href="mailto:business@availproject.org">
+            business@availproject.org
+          </a>
+          .
+        </p>
+      </InfoSection>
+
+      <InfoSection heading="Elsewhere">
+        <p>
+          Announcements and updates are posted on{" "}
+          <a href="https://x.com/AvailProject">X</a>. General information about
+          the project, including the team and open roles, lives on the{" "}
+          <a href="https://www.availproject.org">main Avail website</a>.
+        </p>
+      </InfoSection>
+    </InfoPage>
+  );
+}
diff --git a/src/app/(home)/page.tsx b/src/app/(home)/page.tsx
index de4fdc9..451b5fb 100644
--- a/src/app/(home)/page.tsx
+++ b/src/app/(home)/page.tsx
@@ -3,10 +3,21 @@ import { HeroSection } from "@/components/home/hero-section";
 import { HomeControls } from "@/components/home/home-controls";
 import { ProductGrid } from "@/components/home/product-grid";
 import { SiteFooter } from "@/components/home/site-footer";
+import { SITE_URL, websiteLd } from "@/lib/structured-data";
+
+export const metadata = {
+  alternates: { canonical: SITE_URL },
+};
 
 export default function HomePage() {
   return (
     <main id="main-content" className="relative">
+      {/* Site identity for agents — not rendered */}
+      <script
+        type="application/ld+json"
+        // biome-ignore lint/security/noDangerouslySetInnerHtml: JSON-LD payload is built from static, non-user data
+        dangerouslySetInnerHTML={{ __html: JSON.stringify(websiteLd) }}
+      />
       {/* Frame padding — thin muted border visible on top + sides */}
       <div className="md:px-3 md:pt-3">
         {/* Content panel — white bg, rounded corners */}
diff --git a/src/app/(home)/privacy/page.tsx b/src/app/(home)/privacy/page.tsx
new file mode 100644
index 0000000..235b61a
--- /dev/null
+++ b/src/app/(home)/privacy/page.tsx
@@ -0,0 +1,99 @@
+import { InfoPage, InfoSection } from "@/components/info/info-page";
+import { SITE_URL } from "@/lib/structured-data";
+
+export const metadata = {
+  title: "Privacy",
+  description:
+    "What the Avail documentation site collects — analytics, stored preferences, optional wallet connections, and page feedback — and how to opt out.",
+  alternates: { canonical: `${SITE_URL}/privacy` },
+};
+
+export default function PrivacyPage() {
+  return (
+    <InfoPage
+      title="Privacy"
+      lede="This page describes what docs.availproject.org collects when you read it. It covers this documentation site only — not the Avail network itself, which is a public blockchain, and not other Avail properties."
+    >
+      <InfoSection heading="Analytics">
+        <p>
+          We use PostHog to understand which pages are useful and where people
+          get stuck. It records page views, how far down a page you scroll, how
+          long you stay, and clicks on links, buttons and forms. Analytics
+          requests are proxied through this domain rather than sent to a
+          third-party domain directly.
+        </p>
+        <p>
+          Where session recordings are captured, all text inputs are masked
+          before anything leaves your browser, so what you type is not recorded.
+          Analytics state is kept in your browser&rsquo;s local storage and a
+          cookie.
+        </p>
+        <p>
+          If your browser sends a <code>Do Not Track</code> signal, analytics
+          are disabled automatically. Blocking the analytics script does not
+          affect any documentation content.
+        </p>
+      </InfoSection>
+
+      <InfoSection heading="Preferences stored in your browser">
+        <p>
+          Two settings are saved locally so the site behaves consistently
+          between visits: the network you have selected (mainnet or testnet) and
+          your light or dark theme choice. These stay in your browser and are
+          not sent to us. Clearing site data removes them.
+        </p>
+      </InfoSection>
+
+      <InfoSection heading="Connecting a wallet">
+        <p>
+          Some Avail Nexus pages include live demo widgets that you can connect
+          a wallet to. Connecting is entirely optional, and no documentation
+          requires it. When you connect, the site reads your public address and
+          the balances needed to render the demo. We never receive your private
+          keys or seed phrase, and every transaction is signed in your own
+          wallet.
+        </p>
+      </InfoSection>
+
+      <InfoSection heading="Page feedback">
+        <p>
+          Each page has a feedback control. When you submit it, we receive your
+          rating, the path of the page you were on, and anything you chose to
+          add — a comment, contact details, or a screenshot.
+        </p>
+        <p>
+          Please note that feedback is filed as an issue in a{" "}
+          <a href="https://github.com/availproject/docs">
+            public GitHub repository
+          </a>
+          . Anything you include, including contact details and screenshots,
+          will be publicly visible. Leave out anything you would not want
+          published, and use the channels on the <a href="/contact">Contact</a>{" "}
+          page for private matters.
+        </p>
+      </InfoSection>
+
+      <InfoSection heading="Hosting">
+        <p>
+          The site is hosted on Vercel, which processes standard request data
+          such as IP address and user agent in order to serve pages and protect
+          the service. Uploaded feedback screenshots are stored in Vercel Blob
+          storage.
+        </p>
+      </InfoSection>
+
+      <InfoSection heading="Questions">
+        <p>
+          For questions about this page, or to ask us to remove something you
+          submitted, email{" "}
+          <a href="mailto:info@availproject.org">info@availproject.org</a>.
+          Security issues should go to{" "}
+          <a href="mailto:security@availproject.org">
+            security@availproject.org
+          </a>{" "}
+          instead — see <a href="/contact">Contact</a>.
+        </p>
+      </InfoSection>
+    </InfoPage>
+  );
+}
diff --git a/src/app/api/markdown/[...slug]/route.ts b/src/app/api/markdown/[...slug]/route.ts
index edac7dc..49f06e9 100644
--- a/src/app/api/markdown/[...slug]/route.ts
+++ b/src/app/api/markdown/[...slug]/route.ts
@@ -1,5 +1,6 @@
 import { type NextRequest, NextResponse } from "next/server";
 import { AGENT_HEADERS } from "@/lib/agent-headers";
+import { agentNotFoundResponse } from "@/lib/agent-not-found";
 import { trackAgentRequest } from "@/lib/analytics/agent-tracking";
 import { getPageMarkdownOverride } from "@/lib/llms";
 import { cleanMarkdownForAgents } from "@/lib/markdown-clean";
@@ -11,10 +12,10 @@ function getPageFromSlug(slug: string[]) {
     return directMatch;
   }
 
-  // DA content is canonicalized under /docs/DA/* while many links and
-  // users request /docs/da/*. Support both casings for API consumers.
+  // DA content is canonicalized under /docs/da/* but older links and some
+  // agents request /docs/DA/*. Support both casings for API consumers.
   if (slug.length > 0 && slug[0].toLowerCase() === "da") {
-    return source.getPage(["DA", ...slug.slice(1)]);
+    return source.getPage(["da", ...slug.slice(1)]);
   }
 
   return null;
@@ -28,7 +29,7 @@ export async function GET(
   const page = getPageFromSlug(slug);
 
   if (!page) {
-    return NextResponse.json({ error: "Page not found" }, { status: 404 });
+    return agentNotFoundResponse(`/docs/${slug.join("/")}`);
   }
 
   try {
diff --git a/src/app/api/markdown/markdown-api.test.ts b/src/app/api/markdown/markdown-api.test.ts
index fa1d594..dfc89e7 100644
--- a/src/app/api/markdown/markdown-api.test.ts
+++ b/src/app/api/markdown/markdown-api.test.ts
@@ -1,5 +1,6 @@
 import { NextRequest } from "next/server";
 import { describe, expect, it, vi } from "vitest";
+import { MARKDOWN_VARY } from "@/lib/agent-headers";
 
 const FAKE_CONTENT = "# Test Page\n\nSome markdown content here for testing.";
 // cleanMarkdownForAgents adds a trailing newline to plain markdown
@@ -8,10 +9,10 @@ const CLEANED_CONTENT = `${FAKE_CONTENT}\n`;
 vi.mock("@/lib/source", () => ({
   source: {
     getPage: (slug: string[]) => {
-      if (slug.length === 0 || (slug[0] === "DA" && slug[1] === "build")) {
+      if (slug.length === 0 || (slug[0] === "da" && slug[1] === "build")) {
         return {
           url:
-            slug.length === 0 ? "/docs" : `/docs/DA/${slug.slice(1).join("/")}`,
+            slug.length === 0 ? "/docs" : `/docs/da/${slug.slice(1).join("/")}`,
           data: {
             title: "Test Page",
             description: "A test page",
@@ -62,9 +63,11 @@ describe("markdown API - root route", () => {
     );
   });
 
-  it("has Vary: accept header", async () => {
+  it("has a merged Vary header including Accept", async () => {
     const res = await rootGET(makeRequest("/api/markdown"));
-    expect(res.headers.get("Vary")).toBe("accept");
+    const vary = res.headers.get("Vary") ?? "";
+    expect(vary).toBe(MARKDOWN_VARY);
+    expect(vary).toContain("Accept");
   });
 
   it("returns JSON when format=json", async () => {
diff --git a/src/app/api/markdown/route.ts b/src/app/api/markdown/route.ts
index 445eb80..b02efae 100644
--- a/src/app/api/markdown/route.ts
+++ b/src/app/api/markdown/route.ts
@@ -1,5 +1,6 @@
 import { type NextRequest, NextResponse } from "next/server";
 import { AGENT_HEADERS } from "@/lib/agent-headers";
+import { agentNotFoundResponse } from "@/lib/agent-not-found";
 import { trackAgentRequest } from "@/lib/analytics/agent-tracking";
 import { cleanMarkdownForAgents } from "@/lib/markdown-clean";
 import { source } from "@/lib/source";
@@ -9,7 +10,7 @@ export async function GET(request: NextRequest) {
   const page = source.getPage([]);
 
   if (!page) {
-    return NextResponse.json({ error: "Page not found" }, { status: 404 });
+    return agentNotFoundResponse("/docs");
   }
 
   try {
diff --git a/src/app/docs/[[...slug]]/page.tsx b/src/app/docs/[[...slug]]/page.tsx
index 42e7134..6097827 100644
--- a/src/app/docs/[[...slug]]/page.tsx
+++ b/src/app/docs/[[...slug]]/page.tsx
@@ -11,6 +11,7 @@ import { PageFooter } from "@/components/mdx/page-footer";
 import { Badge } from "@/components/ui/badge";
 import { getProductTree } from "@/lib/page-tree-utils";
 import { source } from "@/lib/source";
+import { organizationLd } from "@/lib/structured-data";
 
 export const revalidate = false;
 export const dynamic = "force-static";
@@ -159,11 +160,15 @@ export default async function Page(props: {
     headline: doc.title,
     description: doc.description || "",
     url: pageUrl,
-    publisher: {
-      "@type": "Organization",
-      name: "Avail",
+    mainEntityOfPage: pageUrl,
+    inLanguage: "en",
+    isPartOf: {
+      "@type": "WebSite",
+      "@id": `${baseUrl}/#website`,
+      name: "Avail Documentation",
       url: baseUrl,
     },
+    publisher: organizationLd,
   };
 
   return (
diff --git a/src/app/not-found.tsx b/src/app/not-found.tsx
index e6840a8..a631bbb 100644
--- a/src/app/not-found.tsx
+++ b/src/app/not-found.tsx
@@ -1,5 +1,6 @@
 import { headers } from "next/headers";
 import { redirect } from "next/navigation";
+import { AGENT_RECOVERY_LINKS } from "@/lib/agent-not-found";
 
 function isAgentRequest(headersList: Headers): boolean {
   const accept = headersList.get("accept") ?? "";
@@ -20,7 +21,22 @@ export default async function NotFound() {
       <html lang="en">
         <body>
           <h1>404 - Page Not Found</h1>
-          <p>This page does not exist. See /llms.txt for a full page index.</p>
+          <p>
+            This page does not exist in the Avail documentation. Use one of the
+            indexes below to find the correct URL.
+          </p>
+          <ul>
+            {AGENT_RECOVERY_LINKS.map(([href, label]) => (
+              <li key={href}>
+                <a href={href}>{href}</a> — {label}
+              </li>
+            ))}
+          </ul>
+          <p>
+            Any <code>/docs/*</code> URL also returns markdown when requested
+            with <code>Accept: text/markdown</code>, or by appending
+            <code>.md</code> to the path.
+          </p>
         </body>
       </html>
     );
diff --git a/src/app/sitemap.test.ts b/src/app/sitemap.test.ts
index 52b5ddf..9506f84 100644
--- a/src/app/sitemap.test.ts
+++ b/src/app/sitemap.test.ts
@@ -15,19 +15,27 @@ import sitemap from "./sitemap";
 describe("sitemap", () => {
   it("returns entries with full URLs", () => {
     const entries = sitemap();
-    expect(entries.length).toBe(3);
     for (const entry of entries) {
-      expect(entry.url).toMatch(/^https:\/\/docs\.availproject\.org\/docs\//);
+      expect(entry.url).toMatch(/^https:\/\/docs\.availproject\.org\//);
     }
   });
 
-  it("sets changeFrequency to weekly", () => {
-    const entries = sitemap();
+  it("sets changeFrequency to weekly for docs pages", () => {
+    const entries = sitemap().filter((e) => e.url.includes("/docs/"));
+    expect(entries.length).toBe(3);
     for (const entry of entries) {
       expect(entry.changeFrequency).toBe("weekly");
     }
   });
 
+  it("includes the homepage and trust anchor pages", () => {
+    const urls = sitemap().map((e) => e.url);
+    expect(urls).toContain("https://docs.availproject.org/");
+    expect(urls).toContain("https://docs.availproject.org/about");
+    expect(urls).toContain("https://docs.availproject.org/contact");
+    expect(urls).toContain("https://docs.availproject.org/privacy");
+  });
+
   it("includes correct page URLs", () => {
     const entries = sitemap();
     const urls = entries.map((e) => e.url);
diff --git a/src/app/sitemap.ts b/src/app/sitemap.ts
index c730210..aa25d70 100644
--- a/src/app/sitemap.ts
+++ b/src/app/sitemap.ts
@@ -3,11 +3,24 @@ import { source } from "@/lib/source";
 
 const BASE_URL = "https://docs.availproject.org";
 
+/**
+ * Routes that exist outside the docs content source. Without these the
+ * homepage and the trust anchor pages are absent from the sitemap, leaving
+ * agents no way to discover who publishes this site.
+ */
+const STATIC_ROUTES = ["/", "/about", "/contact", "/privacy"] as const;
+
 export default function sitemap(): MetadataRoute.Sitemap {
   const pages = source.getPages();
 
-  return pages.map((page) => ({
-    url: `${BASE_URL}${page.url}`,
-    changeFrequency: "weekly" as const,
-  }));
+  return [
+    ...STATIC_ROUTES.map((route) => ({
+      url: `${BASE_URL}${route}`,
+      changeFrequency: "monthly" as const,
+    })),
+    ...pages.map((page) => ({
+      url: `${BASE_URL}${page.url}`,
+      changeFrequency: "weekly" as const,
+    })),
+  ];
 }
diff --git a/src/components/info/info-page.tsx b/src/components/info/info-page.tsx
new file mode 100644
index 0000000..e52a777
--- /dev/null
+++ b/src/components/info/info-page.tsx
@@ -0,0 +1,63 @@
+import Link from "next/link";
+import { SiteFooter } from "@/components/home/site-footer";
+
+/**
+ * Shell for the standalone informational pages (About, Contact, Privacy).
+ *
+ * Mirrors the home page frame — muted ground, bordered content panel, shared
+ * footer — so these pages read as part of the site rather than bolted on.
+ */
+export function InfoPage({
+  title,
+  lede,
+  children,
+}: {
+  title: string;
+  lede: string;
+  children: React.ReactNode;
+}) {
+  return (
+    <main id="main-content" className="relative">
+      <div className="md:px-3 md:pt-3">
+        <div className="relative overflow-hidden md:border md:border-border bg-background">
+          <div className="mx-auto w-full max-w-160 px-6 py-16 md:py-24">
+            <nav className="mb-6">
+              <Link
+                href="/"
+                className="ui-16 text-muted-foreground transition-colors hover:text-foreground"
+              >
+                ← Avail Documentation
+              </Link>
+            </nav>
+            <h1 className="font-serif text-2xl md:text-[28px] font-medium leading-8 md:leading-9 tracking-[0.56px] text-brand">
+              {title}
+            </h1>
+            <p className="body-16 mt-4 text-secondary-foreground">{lede}</p>
+            <div className="mt-10 flex flex-col gap-10">{children}</div>
+          </div>
+        </div>
+      </div>
+      <div className="relative z-10">
+        <SiteFooter />
+      </div>
+    </main>
+  );
+}
+
+/** A titled block within an {@link InfoPage}. */
+export function InfoSection({
+  heading,
+  children,
+}: {
+  heading: string;
+  children: React.ReactNode;
+}) {
+  return (
+    <section className="flex flex-col gap-3">
+      <h2 className="ui-16 font-medium text-foreground">{heading}</h2>
+      <div className="body-16 flex flex-col gap-3 text-secondary-foreground">
+        {children}
+      </div>
+    </section>
+  );
+}
diff --git a/src/lib/agent-headers.ts b/src/lib/agent-headers.ts
index 74c8243..d4ffcc2 100644
--- a/src/lib/agent-headers.ts
+++ b/src/lib/agent-headers.ts
@@ -1,5 +1,15 @@
+/**
+ * Next adds its own `Vary` for RSC negotiation. Emitting a bare
+ * `Vary: Accept` alongside it produces two separate `Vary` header lines,
+ * and CDNs (and agent-readiness scanners) commonly read only the first —
+ * which risks serving an HTML response from a markdown cache entry.
+ * Carry every token in one merged value instead.
+ */
+export const MARKDOWN_VARY =
+  "Accept, RSC, Next-Router-State-Tree, Next-Router-Prefetch, Next-Router-Segment-Prefetch";
+
 export const AGENT_HEADERS = {
   "Content-Signal": "ai-train=yes, search=yes, ai-input=yes",
   "X-Robots-Tag": "noindex",
-  Vary: "accept",
+  Vary: MARKDOWN_VARY,
 };
diff --git a/src/lib/agent-not-found.ts b/src/lib/agent-not-found.ts
new file mode 100644
index 0000000..90fff89
--- /dev/null
+++ b/src/lib/agent-not-found.ts
@@ -0,0 +1,49 @@
+import { NextResponse } from "next/server";
+import { AGENT_HEADERS } from "@/lib/agent-headers";
+
+/**
+ * Recovery pointers for an agent that requested a page that does not exist.
+ * A dead end costs an agent a whole turn, so every 404 names the indexes it
+ * can use to find the right URL instead.
+ */
+export const AGENT_RECOVERY_LINKS = [
+  ["/llms.txt", "Intent-based index of every documented task"],
+  ["/llms-da.txt", "Avail DA documentation (single file)"],
+  ["/llms-nexus.txt", "Avail Nexus documentation (single file)"],
+  ["/llms-full.txt", "Complete documentation dump"],
+  ["/sitemap.xml", "Every canonical page URL"],
+  ["/api/reference.json", "Structured reference data (JSON)"],
+] as const;
+
+function recoveryList(): string {
+  return AGENT_RECOVERY_LINKS.map(
+    ([href, label]) => `- [${href}](${href}): ${label}`,
+  ).join("\n");
+}
+
+/** Markdown 404 body pointing an agent at the discovery indexes. */
+export function agentNotFoundMarkdown(requestedPath?: string): string {
+  const requested = requestedPath ? `\n\nRequested: \`${requestedPath}\`` : "";
+  return `# 404 — Page Not Found
+
+This page does not exist in the Avail documentation.${requested}
+
+## Where to look instead
+
+${recoveryList()}
+
+Any \`/docs/*\` URL also returns markdown when requested with
+\`Accept: text/markdown\`, or by appending \`.md\` to the path.
+`;
+}
+
+/** A 404 whose body is markdown rather than an opaque JSON error. */
+export function agentNotFoundResponse(requestedPath?: string) {
+  return new NextResponse(agentNotFoundMarkdown(requestedPath), {
+    status: 404,
+    headers: {
+      "Content-Type": "text/markdown; charset=utf-8",
+      ...AGENT_HEADERS,
+    },
+  });
+}
diff --git a/src/lib/footer-links.ts b/src/lib/footer-links.ts
index 70b2d0a..26d7e7e 100644
--- a/src/lib/footer-links.ts
+++ b/src/lib/footer-links.ts
@@ -22,6 +22,8 @@ export const pagesColumn1: FooterLink[] = [
 
 export const pagesColumn2: FooterLink[] = [
   { label: "Blog", href: "https://blog.availproject.org", external: true },
+  { label: "About these docs", href: "/about" },
+  { label: "Privacy", href: "/privacy" },
   {
     label: "Ecosystem",
     href: "https://www.availproject.org/ecosystem",
@@ -41,6 +43,7 @@ export const pagesColumn2: FooterLink[] = [
 
 export const supportLinks: FooterLink[] = [
   { label: "Docs", href: "/docs/da/get-started" },
+  { label: "Contact", href: "/contact" },
   {
     label: "Discord",
     href: "https://discord.com/invite/AvailProject",
diff --git a/src/lib/llms.ts b/src/lib/llms.ts
index 5199d47..de8e801 100644
--- a/src/lib/llms.ts
+++ b/src/lib/llms.ts
@@ -71,6 +71,23 @@ export function generateLlmsTxt(): string {
     "| [/llms-nexus.txt](/llms-nexus.txt) | Avail Nexus — cross-chain unification (SDK, concepts, contracts, UI elements) |",
     "| [/llms-full.txt](/llms-full.txt) | Complete documentation dump (all pages) |",
     "",
+    "## Agent Endpoints",
+    "",
+    "Machine-readable entry points for this site:",
+    "",
+    "| Endpoint | Returns |",
+    "|----------|---------|",
+    "| `/llms.txt` | This index |",
+    "| `/llms-da.txt`, `/llms-nexus.txt`, `/llms-full.txt` | Section and full-corpus dumps |",
+    "| `/api/reference.json` | Structured reference data (networks, endpoints, contracts) |",
+    "| `/api/markdown/{slug}` | Clean markdown for one page (add `?format=json` for metadata) |",
+    "| `/docs/{slug}.md` | Same markdown, via URL suffix |",
+    "| `/api/search?query={term}` | Full-text search over all pages |",
+    "| `/sitemap.xml` | Every canonical page URL |",
+    "",
+    "Any `/docs/*` URL returns markdown instead of HTML when requested with",
+    "the header `Accept: text/markdown`.",
+    "",
     "## Quick Reference",
     "",
     "- Mainnet RPC: https://mainnet-rpc.avail.so/rpc",
diff --git a/src/lib/structured-data.ts b/src/lib/structured-data.ts
new file mode 100644
index 0000000..5d1a2f9
--- /dev/null
+++ b/src/lib/structured-data.ts
@@ -0,0 +1,92 @@
+/**
+ * Shared schema.org identity for the docs site.
+ *
+ * These objects are emitted as JSON-LD so agents can establish who publishes
+ * this documentation without scraping the rendered page. Every value here is
+ * sourced from content already present on the site — nothing is asserted that
+ * cannot be verified from an official Avail property.
+ */
+
+export const SITE_URL =
+  process.env.NEXT_PUBLIC_SITE_URL || "https://docs.availproject.org";
+
+const AVAIL_HOME = "https://www.availproject.org";
+
+export const AVAIL_SOCIAL_PROFILES = [
+  "https://x.com/AvailProject",
+  "https://github.com/availproject",
+  "https://discord.com/invite/AvailProject",
+  "https://t.me/AvailCommunity",
+];
+
+export const organizationLd = {
+  "@type": "Organization",
+  "@id": `${AVAIL_HOME}/#organization`,
+  name: "Avail",
+  alternateName: "Avail Project",
+  url: AVAIL_HOME,
+  logo: `${SITE_URL}/avail_logo.png`,
+  description:
+    "Avail is a modular blockchain stack providing data availability (Avail DA) and cross-chain unification (Avail Nexus).",
+  sameAs: AVAIL_SOCIAL_PROFILES,
+  contactPoint: [
+    {
+      "@type": "ContactPoint",
+      contactType: "technical support",
+      url: "https://discord.com/invite/AvailProject",
+      availableLanguage: "en",
+    },
+    {
+      "@type": "ContactPoint",
+      contactType: "sales",
+      email: "business@availproject.org",
+      availableLanguage: "en",
+    },
+    {
+      "@type": "ContactPoint",
+      contactType: "security",
+      email: "security@availproject.org",
+      availableLanguage: "en",
+    },
+  ],
+};
+
+export const websiteLd = {
+  "@context": "https://schema.org",
+  "@graph": [
+    organizationLd,
+    {
+      "@type": "WebSite",
+      "@id": `${SITE_URL}/#website`,
+      url: SITE_URL,
+      name: "Avail Documentation",
+      description:
+        "Documentation for Avail DA (data availability) and Avail Nexus (cross-chain unification).",
+      inLanguage: "en",
+      publisher: { "@id": `${AVAIL_HOME}/#organization` },
+      potentialAction: {
+        "@type": "SearchAction",
+        target: {
+          "@type": "EntryPoint",
+          urlTemplate: `${SITE_URL}/api/search?query={search_term_string}`,
+        },
+        "query-input": "required name=search_term_string",
+      },
+    },
+    {
+      "@type": "SoftwareApplication",
+      "@id": `${SITE_URL}/#software`,
+      name: "Avail",
+      applicationCategory: "DeveloperApplication",
+      operatingSystem: "Linux, macOS, Windows",
+      url: AVAIL_HOME,
+      softwareHelp: SITE_URL,
+      publisher: { "@id": `${AVAIL_HOME}/#organization` },
+      offers: {
+        "@type": "Offer",
+        price: "0",
+        priceCurrency: "USD",
+      },
+    },
+  ],
+};
diff --git a/src/middleware.test.ts b/src/middleware.test.ts
index 0b85e87..0eaf202 100644
--- a/src/middleware.test.ts
+++ b/src/middleware.test.ts
@@ -1,5 +1,6 @@
 import { NextRequest } from "next/server";
 import { describe, expect, it } from "vitest";
+import { MARKDOWN_VARY } from "@/lib/agent-headers";
 import { middleware } from "./middleware";
 
 function makeRequest(
@@ -36,7 +37,7 @@ describe("middleware", () => {
     );
     const rewrite = res.headers.get("x-middleware-rewrite");
     expect(rewrite).not.toBeNull();
-    expect(new URL(rewrite!).pathname).toBe("/api/markdown/DA/build/networks");
+    expect(new URL(rewrite!).pathname).toBe("/api/markdown/da/build/networks");
   });
 
   it("maps root /docs URL to /api/markdown", () => {
@@ -58,13 +59,29 @@ describe("middleware", () => {
       makeRequest("/docs/da/build", "text/markdown", { format: "json" }),
     );
     const rewrite = new URL(res.headers.get("x-middleware-rewrite")!);
-    expect(rewrite.pathname).toBe("/api/markdown/DA/build");
+    expect(rewrite.pathname).toBe("/api/markdown/da/build");
     expect(rewrite.searchParams.get("format")).toBe("json");
   });
 
-  it("sets Vary: accept header", () => {
+  it("sets a single merged Vary header including Accept", () => {
     const res = middleware(makeRequest("/docs/da/build", "text/markdown"));
-    expect(res.headers.get("vary")).toBe("accept");
+    expect(res.headers.get("vary")).toBe(MARKDOWN_VARY);
+  });
+
+  it("preserves DA slug casing rather than uppercasing it", () => {
+    // Regression: the slug used to be rewritten to /api/markdown/DA/*, which
+    // 404'd for every Avail DA page because content is canonicalized as `da`.
+    const res = middleware(
+      makeRequest("/docs/da/get-started", "text/markdown"),
+    );
+    const rewrite = new URL(res.headers.get("x-middleware-rewrite")!);
+    expect(rewrite.pathname).toBe("/api/markdown/da/get-started");
+  });
+
+  it("rewrites .md suffix URLs to the markdown API", () => {
+    const res = middleware(makeRequest("/docs/da/get-started.md"));
+    const rewrite = new URL(res.headers.get("x-middleware-rewrite")!);
+    expect(rewrite.pathname).toBe("/api/markdown/da/get-started");
   });
 
   it("rewrites when text/markdown is among multiple accept types", () => {
@@ -73,6 +90,6 @@ describe("middleware", () => {
     );
     const rewrite = res.headers.get("x-middleware-rewrite");
     expect(rewrite).not.toBeNull();
-    expect(new URL(rewrite!).pathname).toBe("/api/markdown/DA/build");
+    expect(new URL(rewrite!).pathname).toBe("/api/markdown/da/build");
   });
 });
diff --git a/src/middleware.ts b/src/middleware.ts
index aa3be20..fee8555 100644
--- a/src/middleware.ts
+++ b/src/middleware.ts
@@ -1,4 +1,5 @@
 import { type NextRequest, NextResponse } from "next/server";
+import { MARKDOWN_VARY } from "@/lib/agent-headers";
 import { applyRedirects } from "./redirect-middleware";
 
 /**
@@ -12,7 +13,7 @@ import { applyRedirects } from "./redirect-middleware";
  *  - Content-Type: text/markdown
  *  - x-markdown-tokens: estimated token count (chars / 4)
  *  - Content-Signal: ai-train=yes, search=yes, ai-input=yes
- *  - Vary: accept  (so CDN caches both representations)
+ *  - Vary: Accept  (so CDN caches both representations)
  */
 
 export const config = {
@@ -25,19 +26,13 @@ export const config = {
   ],
 };
 
-function normalizeDocsSlugForMarkdownApi(slug: string): string {
-  if (!slug) {
-    return slug;
-  }
-
-  const [section, ...rest] = slug.split("/");
-  // DA docs are canonicalized as /docs/DA/* in the content source.
-  // Normalize lowercase incoming URLs so markdown API lookups succeed.
-  if (section.toLowerCase() === "da") {
-    return ["DA", ...rest].join("/");
-  }
-
-  return slug;
+function rewriteToMarkdown(markdownUrl: URL) {
+  const response = NextResponse.rewrite(markdownUrl);
+  // Set (not append) so the response carries one merged Vary. A separate
+  // `Vary: accept` line is easy for CDNs and scanners to miss, which lets
+  // an HTML response get served from an agent's markdown cache entry.
+  response.headers.set("Vary", MARKDOWN_VARY);
+  return response;
 }
 
 function acceptsMarkdown(request: NextRequest): boolean {
@@ -62,10 +57,9 @@ export function middleware(request: NextRequest) {
   // .md suffix → rewrite to markdown API
   // e.g. /docs/nexus/get-started.md → /api/markdown/nexus/get-started
   if (request.nextUrl.pathname.endsWith(".md")) {
-    const rawSlug = request.nextUrl.pathname
+    const slug = request.nextUrl.pathname
       .replace(/\.md$/, "")
       .replace(/^\/docs\/?/, "");
-    const slug = normalizeDocsSlugForMarkdownApi(rawSlug);
     const markdownUrl = new URL(
       slug ? `/api/markdown/${slug}` : "/api/markdown",
       request.url,
@@ -73,9 +67,7 @@ export function middleware(request: NextRequest) {
     request.nextUrl.searchParams.forEach((value, key) => {
       markdownUrl.searchParams.set(key, value);
     });
-    return NextResponse.rewrite(markdownUrl, {
-      headers: { Vary: "accept" },
-    });
+    return rewriteToMarkdown(markdownUrl);
   }
 
   if (!acceptsMarkdown(request)) {
@@ -83,9 +75,8 @@ export function middleware(request: NextRequest) {
   }
 
   // Strip the /docs prefix to get the slug for the markdown API.
-  // /docs/da/build/networks → /api/markdown/DA/build/networks
-  const rawSlug = request.nextUrl.pathname.replace(/^\/docs\/?/, "");
-  const slug = normalizeDocsSlugForMarkdownApi(rawSlug);
+  // /docs/da/build/networks → /api/markdown/da/build/networks
+  const slug = request.nextUrl.pathname.replace(/^\/docs\/?/, "");
   const markdownUrl = new URL(
     slug ? `/api/markdown/${slug}` : "/api/markdown",
     request.url,
@@ -96,9 +87,5 @@ export function middleware(request: NextRequest) {
     markdownUrl.searchParams.set(key, value);
   });
 
-  return NextResponse.rewrite(markdownUrl, {
-    headers: {
-      Vary: "accept",
-    },
-  });
+  return rewriteToMarkdown(markdownUrl);
 }
diff --git a/src/providers/Web3Provider.tsx b/src/providers/Web3Provider.tsx
index e8cf4e3..3bd721a 100644
--- a/src/providers/Web3Provider.tsx
+++ b/src/providers/Web3Provider.tsx
@@ -130,8 +130,11 @@ const defaultConfig = getDefaultConfig({
 const wagmiConfig = createConfig(defaultConfig);
 
 function NexusContainer({ children }: Readonly<{ children: React.ReactNode }>) {
+  // Render with the default network on the server and on the first client
+  // render so page content is present in the initial HTML. `NexusProvider`
+  // builds its client inside an effect keyed on `config` (destroying the
+  // previous one), so switching to a stored network below is handled safely.
   const [network, setNetwork] = useState<"mainnet" | "testnet">("mainnet");
-  const [isInitialized, setIsInitialized] = useState(false);
 
   useEffect(() => {
     // Initialize network from localStorage on client side
@@ -147,8 +150,6 @@ function NexusContainer({ children }: Readonly<{ children: React.ReactNode }>) {
       setNetwork("mainnet");
       setItem(NETWORK_KEY, "mainnet");
     }
-
-    setIsInitialized(true);
   }, []);
 
   const nexusConfig = useMemo(
@@ -156,11 +157,6 @@ function NexusContainer({ children }: Readonly<{ children: React.ReactNode }>) {
     [network],
   );
 
-  // Don't render until we've initialized from localStorage
-  if (!isInitialized) {
-    return <Skeleton className="w-full h-full" />;
-  }
-
   return (
     <NexusProvider config={nexusConfig}>
       <PostHogIdentify />

From f1804af07372253a20c79d09917371e501aaaa79 Mon Sep 17 00:00:00 2001
From: Prabal <prabal@availproject.org>
Date: Fri, 4 Sep 2026 20:19:47 +0400
Subject: [PATCH 2/3] Fix the two checks the scan report showed were still
 failing
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

Pulled the machine-readable report (is-agentic.com/api/v1/report) for the
production deployment, which gives per-check detail the HTML page omits.
Two things it revealed:

1. The markdown-negotiation check runs against the scanned URL — the site
   root — not a docs page. The middleware matcher covered /docs/* only, so
   `Accept: text/markdown` on / returned HTML and the check failed
   regardless of the DA casing fix. / now negotiates to the docs index.

2. The check reads only the first Vary header. On Vercel's edge Next's
   RSC-only Vary is emitted first, so the merged value was invisible —
   reported as "Vary header missing Accept". Local `next start` happened to
   order these the other way, which is why it looked correct there. Declared
   the merged Vary in vercel.json so the platform value wins.

Redirect rules deliberately do not run for "/" — the matcher covers it
solely for negotiation, and running them would change behaviour for a path
middleware previously never executed on.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
---
 src/middleware.test.ts | 13 +++++++++++++
 src/middleware.ts      | 14 +++++++++++++-
 vercel.json            | 27 +++++++++++++++++++++++++++
 3 files changed, 53 insertions(+), 1 deletion(-)

diff --git a/src/middleware.test.ts b/src/middleware.test.ts
index 0eaf202..70b0a89 100644
--- a/src/middleware.test.ts
+++ b/src/middleware.test.ts
@@ -78,6 +78,19 @@ describe("middleware", () => {
     expect(rewrite.pathname).toBe("/api/markdown/da/get-started");
   });
 
+  it("serves the docs index as markdown for the homepage", () => {
+    // The agentic scanner negotiates on the site root, not a docs URL.
+    const res = middleware(makeRequest("/", "text/markdown"));
+    const rewrite = new URL(res.headers.get("x-middleware-rewrite")!);
+    expect(rewrite.pathname).toBe("/api/markdown");
+    expect(res.headers.get("vary")).toBe(MARKDOWN_VARY);
+  });
+
+  it("leaves the homepage untouched for browsers", () => {
+    const res = middleware(makeRequest("/", "text/html"));
+    expect(res.headers.get("x-middleware-rewrite")).toBeNull();
+  });
+
   it("rewrites .md suffix URLs to the markdown API", () => {
     const res = middleware(makeRequest("/docs/da/get-started.md"));
     const rewrite = new URL(res.headers.get("x-middleware-rewrite")!);
diff --git a/src/middleware.ts b/src/middleware.ts
index fee8555..dcdadd2 100644
--- a/src/middleware.ts
+++ b/src/middleware.ts
@@ -19,6 +19,8 @@ import { applyRedirects } from "./redirect-middleware";
 export const config = {
   // Run on docs URLs (for markdown rewrites) and on legacy paths that need redirects.
   matcher: [
+    // The homepage participates in markdown negotiation only (see below).
+    "/",
     "/docs/:path*",
     "/da/:path*",
     "/nexus/:path*",
@@ -43,7 +45,17 @@ function acceptsMarkdown(request: NextRequest): boolean {
 }
 
 export function middleware(request: NextRequest) {
-  // First, try to apply redirect rules for legacy URLs.
+  // Homepage. Agents asking for the site as markdown get the docs index
+  // rather than an HTML shell. Redirect rules deliberately do not run here:
+  // the matcher covers "/" solely for content negotiation, and running them
+  // would change behaviour for a path middleware previously never saw.
+  if (request.nextUrl.pathname === "/") {
+    return acceptsMarkdown(request)
+      ? rewriteToMarkdown(new URL("/api/markdown", request.url))
+      : NextResponse.next();
+  }
+
+  // Then, try to apply redirect rules for legacy URLs.
   const redirectResponse = applyRedirects(request);
   if (redirectResponse) {
     return redirectResponse;
diff --git a/vercel.json b/vercel.json
index cbf4d7d..f9850ab 100644
--- a/vercel.json
+++ b/vercel.json
@@ -1,5 +1,32 @@
 {
   "headers": [
+    {
+      "source": "/",
+      "headers": [
+        {
+          "key": "Vary",
+          "value": "Accept, RSC, Next-Router-State-Tree, Next-Router-Prefetch, Next-Router-Segment-Prefetch"
+        }
+      ]
+    },
+    {
+      "source": "/docs/:path*",
+      "headers": [
+        {
+          "key": "Vary",
+          "value": "Accept, RSC, Next-Router-State-Tree, Next-Router-Prefetch, Next-Router-Segment-Prefetch"
+        }
+      ]
+    },
+    {
+      "source": "/api/markdown/:path*",
+      "headers": [
+        {
+          "key": "Vary",
+          "value": "Accept, RSC, Next-Router-State-Tree, Next-Router-Prefetch, Next-Router-Segment-Prefetch"
+        }
+      ]
+    },
     {
       "source": "/fonts/(.*)",
       "headers": [

From 83c0312d54bbdd84b856cd480413193fe871b13f Mon Sep 17 00:00:00 2001
From: Prabal <prabal@availproject.org>
Date: Fri, 4 Sep 2026 20:30:14 +0400
Subject: [PATCH 3/3] =?UTF-8?q?Revert=20the=20vercel.json=20Vary=20rules?=
 =?UTF-8?q?=20=E2=80=94=20they=20have=20no=20effect?=
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

Measured on the preview: header rules in vercel.json are not applied to
Next-served routes. HTML responses for both / and /docs/* came back with
only Next's own `vary: rsc, next-router-state-tree, ...`, so the rules
were dead config. The equivalent in next.config.mjs headers() was also
tested and appends rather than replaces, leaving Next's RSC-only Vary
first, so it is not kept either.

Accept still reaches the client as a second Vary line via AGENT_HEADERS,
which is valid HTTP — RFC 9110 treats repeated field lines as equivalent
to one comma-joined value. Whether the scanner credits it depends on
whether it combines duplicate headers or reads only the first.

The homepage negotiation fix from the previous commit is unaffected and
verified working on the preview: / with Accept: text/markdown now returns
text/markdown, and browsers still get HTML.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
---
 vercel.json | 27 ---------------------------
 1 file changed, 27 deletions(-)

diff --git a/vercel.json b/vercel.json
index f9850ab..cbf4d7d 100644
--- a/vercel.json
+++ b/vercel.json
@@ -1,32 +1,5 @@
 {
   "headers": [
-    {
-      "source": "/",
-      "headers": [
-        {
-          "key": "Vary",
-          "value": "Accept, RSC, Next-Router-State-Tree, Next-Router-Prefetch, Next-Router-Segment-Prefetch"
-        }
-      ]
-    },
-    {
-      "source": "/docs/:path*",
-      "headers": [
-        {
-          "key": "Vary",
-          "value": "Accept, RSC, Next-Router-State-Tree, Next-Router-Prefetch, Next-Router-Segment-Prefetch"
-        }
-      ]
-    },
-    {
-      "source": "/api/markdown/:path*",
-      "headers": [
-        {
-          "key": "Vary",
-          "value": "Accept, RSC, Next-Router-State-Tree, Next-Router-Prefetch, Next-Router-Segment-Prefetch"
-        }
-      ]
-    },
     {
       "source": "/fonts/(.*)",
       "headers": [