diff --git a/.github/workflows/check-llms.yml b/.github/workflows/check-llms.yml
new file mode 100644
index 000000000..bcc37facc
--- /dev/null
+++ b/.github/workflows/check-llms.yml
@@ -0,0 +1,39 @@
+name: Check llms.txt
+
+# Guards against llms.txt drifting out of sync with the site product menu.
+# Runs the dependency-free scripts/check-llms-links.mjs; fails the check if a
+# product in documents.ts has no matching link in llms.txt. Kept separate from
+# the deploy workflow so it never blocks a production deploy.
+
+on:
+  pull_request:
+    paths:
+      - "docs/.vuepress/public/llms.txt"
+      - "docs/.vuepress/config-client/documents.ts"
+      - "scripts/check-llms-links.mjs"
+      - ".github/workflows/check-llms.yml"
+  push:
+    branches:
+      - master
+    paths:
+      - "docs/.vuepress/public/llms.txt"
+      - "docs/.vuepress/config-client/documents.ts"
+      - "scripts/check-llms-links.mjs"
+
+jobs:
+  check-llms:
+    runs-on: ubuntu-latest
+    timeout-minutes: 5
+    steps:
+      - name: Checkout
+        uses: actions/checkout@v4
+        with:
+          persist-credentials: false
+
+      - name: Setup Node.js
+        uses: actions/setup-node@v4
+        with:
+          node-version: "22"
+
+      - name: Check llms.txt is in sync with the product menu
+        run: node scripts/check-llms-links.mjs
diff --git a/.gitignore b/.gitignore
index 5b7fc77c3..28d363076 100644
--- a/.gitignore
+++ b/.gitignore
@@ -4,3 +4,6 @@ docs/.vuepress/dist
 docs/.vuepress/.cache
 docs/.vuepress/.temp
 package-lock.json
+
+# Generated at build time by scripts/generate-llms-full.mjs
+docs/.vuepress/public/llms-full.txt
diff --git a/.cursor/mcp.json b/.mcp.json
similarity index 100%
rename from .cursor/mcp.json
rename to .mcp.json
diff --git a/docs/.vuepress/public/llms.txt b/docs/.vuepress/public/llms.txt
new file mode 100644
index 000000000..1569cba4a
--- /dev/null
+++ b/docs/.vuepress/public/llms.txt
@@ -0,0 +1,41 @@
+# TuxCare Documentation
+
+> TuxCare provides automated Linux patch management, live kernel patching, extended lifecycle support (ELS) for end-of-life operating systems, libraries, runtimes and applications, and SecureChain — verified, signed, continuously patched open-source packages. This site (docs.tuxcare.com) is the official product documentation. Each link below points to a product area; append the path to https://docs.tuxcare.com to read it.
+
+The full text of every documentation page, concatenated into a single file, is available at [https://docs.tuxcare.com/llms-full.txt](https://docs.tuxcare.com/llms-full.txt).
+
+## Products
+
+- [TuxCare](https://docs.tuxcare.com/tuxcare/): Overview of TuxCare services that automate, simplify, and enhance Linux operations.
+- [TuxCare Radar](https://docs.tuxcare.com/radar/): Vulnerability risk assessment and prioritization that reveals real-world risk beyond conventional scanner scores.
+- [TuxCare Enterprise Support](https://docs.tuxcare.com/enterprise-support-for-almalinux/): Extended support and security fixes for AlmaLinux and Rocky Linux.
+
+## Live patching
+
+- [KernelCare](https://docs.tuxcare.com/live-patching-services/): Live kernel patching — security patches and bugfixes with no reboot.
+- [KernelCare for IoT](https://docs.tuxcare.com/kernelcare-for-iot/): Live patching for ARM64 embedded/IoT systems.
+- [ePortal](https://docs.tuxcare.com/eportal/): Web management console for KernelCare Enterprise live patch management.
+- [ePortal API](https://docs.tuxcare.com/eportal-api/): API reference for patch management and ePortal automation.
+
+## Extended Lifecycle Support (ELS)
+
+- [ELS for Operating Systems](https://docs.tuxcare.com/els-for-os/): Post-EOL security support for end-of-life Linux distributions.
+- [ELS for Libraries](https://docs.tuxcare.com/els-for-libraries/): Extended security fixes for open-source libraries.
+- [ELS for Runtimes](https://docs.tuxcare.com/els-for-runtimes/): Extended support for language runtimes (PHP, Python, Node.js, etc.).
+- [ELS for Applications](https://docs.tuxcare.com/els-for-applications/): Security patches for end-of-life open-source applications.
+
+## SecureChain
+
+- [SecureChain for Open Source](https://docs.tuxcare.com/securechain/): Verified, signed, continuously patched open-source packages from a TuxCare-managed registry — drop-in replacements that extend coverage beyond upstream EOL.
+- [SecureChain JavaScript Ecosystem Guide](https://docs.tuxcare.com/securechain/javascript/): Connect to the SecureChain npm registry, authenticate, install dependencies, and troubleshoot resolution.
+- [Managing the SecureChain Repository](https://docs.tuxcare.com/securechain/managing-securechain-repository/): Upgrade installed SecureChain packages to newer releases across ecosystems.
+
+## Account & reference
+
+- [Subscription Management Portal](https://docs.tuxcare.com/tuxcare-cln/): Manage licenses and subscriptions for TuxCare products and services.
+- [Service Descriptions](https://docs.tuxcare.com/service-descriptions/): Service-level and technical support details.
+
+## Optional
+
+- [TuxCare product site](https://tuxcare.com/): Marketing/product overview and contact.
+- [CVE Tracker](https://tuxcare.com/cve-tracker/): Track vulnerability fixes and update status across TuxCare-supported packages.
diff --git a/package.json b/package.json
index fe08cf4a1..e5edb67b6 100644
--- a/package.json
+++ b/package.json
@@ -5,7 +5,7 @@
   "type": "module",
   "scripts": {
     "docs:dev": "vuepress dev docs",
-    "docs:build": "vuepress build docs",
+    "docs:build": "node scripts/generate-llms-full.mjs && vuepress build docs",
     "dev": "vite",
     "build": "vue-tsc && vite build",
     "preview": "vite preview"
diff --git a/scripts/check-llms-links.mjs b/scripts/check-llms-links.mjs
new file mode 100644
index 000000000..b90e8a6da
--- /dev/null
+++ b/scripts/check-llms-links.mjs
@@ -0,0 +1,86 @@
+#!/usr/bin/env node
+/**
+ * Drift guard for llms.txt.
+ *
+ * The curated llms.txt is hand-maintained. Its one real staleness risk is a
+ * product being added to (or renamed in) the site's product menu without a
+ * matching entry in llms.txt. This check compares the product links in
+ * docs/.vuepress/config-client/documents.ts (the source of truth for the
+ * product menu) against llms.txt and fails if any product is missing.
+ *
+ * It intentionally does NOT require the reverse: llms.txt may include extra
+ * links (SecureChain subsections, the Optional marketing/CVE-tracker links)
+ * that are not products in documents.ts.
+ *
+ * No dependencies — runs on plain Node built-ins.
+ */
+import { existsSync, readFileSync } from "node:fs";
+import { dirname, join, resolve } from "node:path";
+import { fileURLToPath } from "node:url";
+
+const REPO = resolve(dirname(fileURLToPath(import.meta.url)), "..");
+const DOCS = join(REPO, "docs");
+const DOCUMENTS = join(DOCS, ".vuepress", "config-client", "documents.ts");
+const LLMS = join(DOCS, ".vuepress", "public", "llms.txt");
+const SITE = "https://docs.tuxcare.com";
+
+const documentsSrc = readFileSync(DOCUMENTS, "utf-8");
+const llms = readFileSync(LLMS, "utf-8");
+
+// Extract each product's title + link from the documents.ts array literal.
+const productRe = /title:\s*"([^"]+)"[\s\S]*?link:\s*"([^"]+)"/g;
+const products = [...documentsSrc.matchAll(productRe)].map((m) => ({
+  title: m[1],
+  link: m[2],
+}));
+
+if (products.length === 0) {
+  console.error(
+    `[check-llms-links] Could not parse any products from ${DOCUMENTS}. ` +
+      `Has the file format changed?`,
+  );
+  process.exit(1);
+}
+
+const missing = products.filter(({ link }) => !llms.includes(SITE + link));
+
+if (missing.length > 0) {
+  console.error(
+    "[check-llms-links] llms.txt is out of sync with the site product menu.\n" +
+      "The following products in docs/.vuepress/config-client/documents.ts " +
+      "have no matching link in docs/.vuepress/public/llms.txt:\n",
+  );
+  for (const { title, link } of missing) {
+    console.error(`  - ${title}  ->  ${SITE + link}`);
+  }
+  console.error(
+    "\nAdd each missing product to llms.txt (with a curated one-line " +
+      "description), then re-run this check.",
+  );
+  process.exit(1);
+}
+
+// Validity: every docs.tuxcare.com page link in llms.txt must resolve to a
+// real page on disk. Catches typos and links left stale when a product is
+// renamed or removed in documents.ts. Page URLs end in "/"; the llms-full.txt
+// file link (and any other non-page link) is skipped.
+const pageRe = /https:\/\/docs\.tuxcare\.com(\/[\w/-]*\/)/g;
+const broken = [...new Set([...llms.matchAll(pageRe)].map((m) => m[1]))].filter(
+  (path) => !existsSync(join(DOCS, path, "README.md")),
+);
+
+if (broken.length > 0) {
+  console.error(
+    "[check-llms-links] llms.txt contains links to pages that do not exist " +
+      "on disk (typo, or a product renamed/removed in documents.ts?):\n",
+  );
+  for (const path of broken) {
+    console.error(`  - ${SITE + path}  ->  docs${path}README.md (not found)`);
+  }
+  process.exit(1);
+}
+
+console.log(
+  `[check-llms-links] OK — all ${products.length} product links present, ` +
+    `every page link resolves on disk.`,
+);
diff --git a/scripts/generate-llms-full.mjs b/scripts/generate-llms-full.mjs
new file mode 100644
index 000000000..542cf0253
--- /dev/null
+++ b/scripts/generate-llms-full.mjs
@@ -0,0 +1,123 @@
+#!/usr/bin/env node
+/**
+ * Generate docs/.vuepress/public/llms-full.txt from the documentation source.
+ *
+ * Runs before `vuepress build` (see the `docs:build` script in package.json),
+ * so the file is always regenerated from the current docs and copied to the
+ * site root by VuePress (public/ -> dist/). The file is git-ignored; it exists
+ * only in the build output.
+ *
+ * Each page is emitted under a `Source:` header with its published URL.
+ * Cleaning applied:
+ *   - strip YAML frontmatter
+ *   - strip HTML comments
+ *   - strip <script>/<style> blocks
+ *   - strip CamelCase VuePress component tags (keep inner prose; salvage
+ *     literal code snippets from CodeTabs/TableTabs `content:` attributes)
+ *   - strip VuePress container markers (:::tip / :::warning / :::), keep body
+ *   - preserve ALL-CAPS placeholder tags (<YOUR_TOKEN>, <EOF>, ...)
+ *   - collapse runs of blank lines
+ */
+import { readFileSync, writeFileSync, readdirSync, statSync } from "node:fs";
+import { dirname, join, relative, resolve } from "node:path";
+import { fileURLToPath } from "node:url";
+
+const __dirname = dirname(fileURLToPath(import.meta.url));
+const REPO = resolve(__dirname, "..");
+const DOCS = join(REPO, "docs");
+const SITE = "https://docs.tuxcare.com";
+const OUT = join(DOCS, ".vuepress", "public", "llms-full.txt");
+const EXCLUDE_DIRS = new Set([".vuepress", "_pages-backup", "node_modules"]);
+
+// CamelCase component name: starts uppercase, contains at least one lowercase.
+const COMP = "[A-Z][A-Za-z0-9]*[a-z][A-Za-z0-9]*";
+const frontmatterRe = /^---\n[\s\S]*?\n---\n/;
+const commentRe = /<!--[\s\S]*?-->/g;
+const scriptRe = /<script\b[^>]*>[\s\S]*?<\/script>/gi;
+const styleRe = /<style\b[^>]*>[\s\S]*?<\/style>/gi;
+const selfCloseRe = new RegExp(`<${COMP}\\b[^>]*?/>`, "g");
+const openRe = new RegExp(`<${COMP}\\b[^>]*?>`, "g");
+const closeRe = new RegExp(`</${COMP}\\s*>`, "g");
+const templateRe = /<\/?template\b[^>]*>/g;
+// A CodeTabs/TableTabs `content:` is either an inline `code` literal or a
+// variable defined in the page's <script setup>; resolve both.
+const scriptVarRe = /(?:const|let|var)\s+([A-Za-z_$][\w$]*)\s*=\s*`([^`]*)`/g;
+const tabRe = /title:\s*'([^']+)'\s*,\s*content:\s*(`[^`]*`|[A-Za-z_$][\w$]*)/g;
+// VuePress container markers: an opening `:::tip Title`, a bare `:::` fence,
+// and custom-container variants. Match the marker line only, keep the body.
+const containerRe = /^[ \t]*:::+.*$/gm;
+
+/** Emit each CodeTabs/TableTabs tab as a titled code block, resolving
+ *  `content:` variable refs against the page's <script> definitions. */
+function salvage(tagText, vars) {
+  const out = [];
+  for (const [, title, ref] of tagText.matchAll(tabRe)) {
+    const code = ref.startsWith("`") ? ref.slice(1, -1) : vars.get(ref);
+    if (code == null) continue;
+    out.push(`${title}:`, "```\n" + code.trim() + "\n```");
+  }
+  return out.join("\n");
+}
+
+function clean(text) {
+  // Capture <script> variable bodies before the script block is stripped.
+  const vars = new Map([...text.matchAll(scriptVarRe)].map((m) => [m[1], m[2]]));
+  text = text.replace(frontmatterRe, "");
+  text = text.replace(commentRe, "");
+  text = text.replace(scriptRe, "");
+  text = text.replace(styleRe, "");
+  text = text.replace(selfCloseRe, (m) => salvage(m, vars));
+  text = text.replace(openRe, "");
+  text = text.replace(closeRe, "");
+  text = text.replace(templateRe, "");
+  text = text.replace(containerRe, "");
+  text = text.replace(/\n{3,}/g, "\n\n");
+  return text.trim();
+}
+
+function urlFor(relPath) {
+  const d = dirname(relPath);
+  return SITE + "/" + (d === "." ? "" : d.replace(/\\/g, "/") + "/");
+}
+
+/** Recursively collect README.md files, skipping excluded/dot directories. */
+function collect(dir, acc = []) {
+  for (const entry of readdirSync(dir)) {
+    const full = join(dir, entry);
+    if (statSync(full).isDirectory()) {
+      if (EXCLUDE_DIRS.has(entry) || entry.startsWith(".")) continue;
+      collect(full, acc);
+    } else if (entry === "README.md") {
+      acc.push(full);
+    }
+  }
+  return acc;
+}
+
+const pages = collect(DOCS)
+  .map((full) => ({ url: urlFor(relative(DOCS, full)), path: full }))
+  .sort((a, b) => a.url.localeCompare(b.url));
+
+const parts = [
+  "# TuxCare Documentation — Full Text",
+  "",
+  "> Full text of the TuxCare product documentation (docs.tuxcare.com), " +
+    "concatenated for LLM consumption. Each section below is one documentation " +
+    "page, preceded by its source URL. Interactive/tabbed content has been " +
+    "flattened; visit the source URL for the rendered page.",
+  "",
+];
+
+let included = 0;
+for (const { url, path } of pages) {
+  const body = clean(readFileSync(path, "utf-8"));
+  if (!body) continue;
+  included += 1;
+  parts.push("\n\n" + "=".repeat(80));
+  parts.push(`Source: ${url}`);
+  parts.push("=".repeat(80) + "\n");
+  parts.push(body);
+}
+
+writeFileSync(OUT, parts.join("\n") + "\n");
+console.log(`[generate-llms-full] pages: ${included}, bytes: ${readFileSync(OUT).length}`);