@lynxflow/seo-engine 1.5.5 → 1.5.7

package/src/index.ts

@@ -15,6 +15,8 @@ export * from "./slug-engine";
 export * from "./legal-disclaimers";
 export * from "./matrix-engine";
 export * from "./brand-icons";
+export * from "./urlytics-engine";
+export * from "./keyword-permutator";
 export * from "./llm-prompt";
 export * from "./auth-key";
 export * from "./token-quota-manager";

package/src/keyword-permutator.ts

@@ -0,0 +1,118 @@
+/**
+ * ⚡ High-Speed Keyword Permutator & SEM Matrix Engine
+ * Inspired by advertools.kw_generate
+ *
+ * Generates all permutations and combinations of Products × Modifiers × Intents × Locations
+ * with match types (Broad, Phrase, Exact) and estimated search intent tagging.
+ */
+
+export type SearchIntentType = "commercial" | "transactional" | "informational" | "navigational";
+
+export interface GeneratedKeyword {
+  keyword: string;
+  exactMatch: string;
+  phraseMatch: string;
+  intent: SearchIntentType;
+  product: string;
+  modifier?: string;
+  location?: string;
+}
+
+export interface PermutatorOptions {
+  products: string[];
+  words?: string[];
+  locations?: string[];
+  maxCombinations?: number;
+}
+
+export class KeywordPermutatorEngine {
+  /**
+   * Generates a combinatorial matrix of keywords.
+   */
+  static generateKeywordMatrix(options: PermutatorOptions): GeneratedKeyword[] {
+    const { products, words = [], locations = [], maxCombinations = 5000 } = options;
+    const results: GeneratedKeyword[] = [];
+
+    const intentKeywords: Record<SearchIntentType, string[]> = {
+      transactional: ["buy", "pricing", "cost", "hire", "quote", "tarif", "prix", "devis", "acheter"],
+      commercial: ["best", "top", "review", "vs", "comparison", "alternative", "comparatif", "meilleur"],
+      informational: ["how to", "what is", "guide", "tutorial", "definition", "comment", "quest ce que"],
+      navigational: ["login", "app", "portal", "website", "connexion"],
+    };
+
+    const detectIntent = (text: string): SearchIntentType => {
+      const lower = text.toLowerCase();
+      for (const [intent, triggers] of Object.entries(intentKeywords) as [SearchIntentType, string[]][]) {
+        if (triggers.some((t) => lower.includes(t))) {
+          return intent;
+        }
+      }
+      return "commercial";
+    };
+
+    for (const prod of products) {
+      // 1. Product alone
+      results.push({
+        keyword: prod,
+        exactMatch: `[${prod}]`,
+        phraseMatch: `"${prod}"`,
+        intent: detectIntent(prod),
+        product: prod,
+      });
+
+      // 2. Product × Words / Modifiers
+      for (const word of words) {
+        const kw1 = `${word} ${prod}`;
+        const kw2 = `${prod} ${word}`;
+
+        results.push({
+          keyword: kw1,
+          exactMatch: `[${kw1}]`,
+          phraseMatch: `"${kw1}"`,
+          intent: detectIntent(word),
+          product: prod,
+          modifier: word,
+        });
+
+        results.push({
+          keyword: kw2,
+          exactMatch: `[${kw2}]`,
+          phraseMatch: `"${kw2}"`,
+          intent: detectIntent(word),
+          product: prod,
+          modifier: word,
+        });
+
+        // 3. Product × Words × Locations
+        for (const loc of locations) {
+          const kwLoc1 = `${word} ${prod} ${loc}`;
+          const kwLoc2 = `${prod} ${loc} ${word}`;
+
+          results.push({
+            keyword: kwLoc1,
+            exactMatch: `[${kwLoc1}]`,
+            phraseMatch: `"${kwLoc1}"`,
+            intent: detectIntent(word),
+            product: prod,
+            modifier: word,
+            location: loc,
+          });
+
+          results.push({
+            keyword: kwLoc2,
+            exactMatch: `[${kwLoc2}]`,
+            phraseMatch: `"${kwLoc2}"`,
+            intent: detectIntent(word),
+            product: prod,
+            modifier: word,
+            location: loc,
+          });
+
+          if (results.length >= maxCombinations) return results;
+        }
+      }
+    }
+
+    return results;
+  }
+}

package/src/llm-prompt.ts

@@ -37,7 +37,21 @@ When discovering product modules, features, or services to build the SEO matrice
 
 ---
 
-### 🏆 3. THE ARCHITECTURAL PLAYBOOK OF SAAS TITANS (7M+ PAGES WITH 0 DB BOTTLENECKS)
+### 📏 3. THE GOLDEN RULES OF URL ARCHITECTURE (MAX 3-4 SEGMENTS, ZERO PARASITE WORDS)
+
+- **Rule A (2 to 3 URL Segments Ideal, 4 Maximum):**
+  - ✅ **2 Segments (Top SEO Performance):** \`site.com/{service}/{city}\` (e.g. \`/autopost-facebook/lyon\`, \`/crm/paris\`).
+  - ✅ **3 Segments (B2B Persona/Hub):** \`site.com/for/{target}/{city}\` or \`site.com/{locale}/{service}/{city}\`.
+  - ❌ **Never 5 to 6 Segments:** Avoid \`/solutions/ai/autopost/facebook/fr/lyon\`.
+- **Rule B (Eliminate All Parasite Noise Words):**
+  - Banish generic wrapper words like \`/solutions/\`, \`/pages/\`, \`/items/\`. Go direct to user search intent: What you do + Platform + Location.
+- **Rule C (The SDK Does 100% of the Heavy Lifting):**
+  - The developer/AI only registers 3-4 product features and target cities in \`lib/seo.ts\`.
+  - The SDK automatically resolves routes in memory via \`matrixEngine.resolvePage(params.slug)\`, generates complete Schema.org graphs (LocalBusiness, AggregateRating, AggregateOffer), creates AEO summaries, builds FAQ accordions, and manages mesh linking in < 0.05ms.
+
+---
+
+### 🏆 4. THE ARCHITECTURAL PLAYBOOK OF SAAS TITANS (7M+ PAGES WITH 0 DB BOTTLENECKS)
 
 How category leaders generate millions of high-converting pages without server saturation:
 
@@ -49,18 +63,18 @@ How category leaders generate millions of high-converting pages without server s
 
 2. **The 4 Pillars of Zero-Latency (< 0.05ms) Resolution:**
    - **0 Bytes in Database:** Never generate physical static files or query heavy databases per request. Pages are resolved mathematically in RAM upon arrival.
-   - **1 Single Universal Catch-All Route:** \`app/[...slug]/page.tsx\` (or \`app/[locale]/[...slug]/page.tsx\`) handles all matrices through one lightweight layout.
-   - **Edge Caching & ISR:** The rendered HTML is instantly cached at the CDN/Edge (Cloudflare / Vercel Edge) on first request.
+   - **1 Single Universal Catch-All Route:** \`app/[...slug]/page.tsx\` handles all matrices through one lightweight layout.
+   - **Edge Caching & ISR:** The rendered HTML is instantly cached at the CDN/Edge on first request.
    - **Automatic Siloing & Mesh Linking:**
      - Links to 3 neighboring geographic cities for the same service.
      - Links to 4 related professions/industries in the same target category.
      - Complete Schema.org BreadcrumbList navigation.
-   - **Sharded XML Sitemaps:** Split into 50,000 URL chunks with a master sitemap index to respect Google's crawl budget.
+   - **Sharded XML Sitemaps:** Split into 50,000 URL chunks with a master sitemap index.
    - **Dynamic \`/llms.txt\` Feed:** High-density Markdown stream enabling direct indexing by ChatGPT Search & Perplexity.
 
 ---
 
-### 🛠️ 4. AI AGENT IMPLEMENTATION PROTOCOL (STEP-BY-STEP)
+### 🛠️ 5. AI AGENT IMPLEMENTATION PROTOCOL (STEP-BY-STEP)
 
 When setting up programmatic SEO in any codebase, the AI Agent MUST generate these exact 4 files:
 
@@ -73,6 +87,7 @@ export const matrixEngine = new PseoMatrixEngine();
 export const SEO_CONFIG = {
   brandName: "Acme",
   language: "en",
+  cleanDirectRoutes: true, // Eliminates parasite words (/solutions/, etc.)
   minPopulationToIndex: 15_000,
 };
 
@@ -80,19 +95,23 @@ export const SEO_CONFIG = {
 export const PSEO_DATASET = {
   services: [
     // 1 per major frontend feature branch
-    { slug: "crm-pipeline", name: "Visual CRM Pipeline", category: "Sales", keyFeatures: ["Kanban", "Lead scoring"], description: "Visual deal tracking" },
+    { slug: "autopost-facebook", name: "Autopost Facebook & Instagram", category: "Social Media", keyFeatures: ["AI Scheduling", "Visual Calendar", "Auto-Hashtags"], description: "Automated social media posting engine" },
+  ],
+  locations: [
+    { slug: "paris", name: "Paris", region: "Île-de-France", country: "France", population: 2_161_000, currency: "EUR", currencySymbol: "€" },
+    { slug: "lyon", name: "Lyon", region: "Rhône-Alpes", country: "France", population: 522_000, currency: "EUR", currencySymbol: "€" },
   ],
   targets: [
-    { slug: "law-firms", name: "Law Firms", type: "industry", painPoints: ["Billing"], benefits: ["Compliance"] },
+    { slug: "agences", name: "Marketing Agencies", type: "industry", painPoints: ["Manual Posting"], benefits: ["10x Output"] },
   ],
   integrations: [
-    { slug: "shopify", name: "Shopify", category: "E-Commerce", syncFeatures: ["Order sync"] },
+    { slug: "shopify", name: "Shopify", category: "E-Commerce", syncFeatures: ["Catalog sync"] },
   ],
   templates: [
-    { slug: "sales-tracker-excel", title: "Sales Tracker", topic: "Sales", format: "excel", benefits: ["Instant setup"] },
+    { slug: "social-calendar-excel", title: "Social Media Calendar", topic: "Social Media", format: "excel", benefits: ["Instant setup"] },
   ],
   glossaryTerms: [
-    { slug: "mrr", term: "MRR", shortDefinition: "Monthly Recurring Revenue" },
+    { slug: "engagement-rate", term: "Engagement Rate", shortDefinition: "Percentage of audience interacting with content" },
   ],
 };
 \`\`\`
@@ -105,19 +124,15 @@ import { matrixEngine, SEO_CONFIG, PSEO_DATASET } from "@/lib/seo";
 import { renderBrandIconSvg } from "@lynxflow/seo-engine";
 
 export async function generateMetadata({ params }: { params: { slug: string[] } }) {
-  const path = "/" + params.slug.join("/");
   const domain = process.env.NEXT_PUBLIC_SITE_URL || "https://example.com";
-  const pages = matrixEngine.generateAllMatrices(domain, PSEO_DATASET, SEO_CONFIG);
-  const page = pages.find((p) => p.urlPath === path);
+  const page = matrixEngine.resolvePage(params.slug, domain, PSEO_DATASET, SEO_CONFIG);
   if (!page) return {};
   return { title: page.title, description: page.description, alternates: { canonical: page.canonicalUrl }, robots: page.robots };
 }
 
 export default async function ProgrammaticPage({ params }: { params: { slug: string[] } }) {
-  const path = "/" + params.slug.join("/");
   const domain = process.env.NEXT_PUBLIC_SITE_URL || "https://example.com";
-  const pages = matrixEngine.generateAllMatrices(domain, PSEO_DATASET, SEO_CONFIG);
-  const page = pages.find((p) => p.urlPath === path);
+  const page = matrixEngine.resolvePage(params.slug, domain, PSEO_DATASET, SEO_CONFIG);
   if (!page) notFound();
 
   // Automatic Brand Icon Detection

package/src/matrix-engine.ts

@@ -155,6 +155,7 @@ export interface GeneratedPageMeta {
 export interface MatrixOptions {
   brandName: string;
   language?: string; // Default: "en", supports "fr", "de", "es", "pt", "it", etc.
+  cleanDirectRoutes?: boolean; // When true (default), eliminates parasite words (/solutions/, etc.) for direct /{service}/{city}
   minPopulationToIndex?: number;
   defaultCurrency?: string;
   defaultCurrencySymbol?: string;
@@ -263,7 +264,10 @@ export class PseoMatrixEngine {
         const citySlug = cleanSeoSlug(loc.slug, { language: lang });
         const serviceSlug = cleanSeoSlug(s.slug, { language: lang });
 
-        const urlPath = `${prefix}/${serviceSlug}/${countryCode}/${citySlug}`;
+        // Zero parasite words: Direct /{service}/{city} or /{service}/{country}/{city} if specified
+        const urlPath = options.cleanDirectRoutes !== false
+          ? `/${serviceSlug}/${citySlug}`
+          : `${prefix}/${serviceSlug}/${countryCode}/${citySlug}`;
         const fullUrl = `${cleanDomain}${urlPath}`;
 
         const isIndexed = (loc.population ?? 20_000) >= minPop;
@@ -815,4 +819,83 @@ export class PseoMatrixEngine {
 
     return allPages;
   }
+
+  /**
+   * Resolves a single page route in RAM in < 0.05ms.
+   * Matches string path or Next.js slug array.
+   */
+  resolvePage(
+    slugOrPath: string | string[],
+    domain: string,
+    data: {
+      services?: PseoService[];
+      locations?: PseoLocation[];
+      competitors?: PseoCompetitor[];
+      targets?: PseoTarget[];
+      integrations?: PseoIntegration[];
+      useCases?: PseoUseCase[];
+      templates?: PseoTemplate[];
+      glossaryTerms?: PseoGlossaryTerm[];
+      calculators?: PseoCalculator[];
+    },
+    options: MatrixOptions,
+  ): GeneratedPageMeta | undefined {
+    const path = typeof slugOrPath === "string"
+      ? (slugOrPath.startsWith("/") ? slugOrPath : `/${slugOrPath}`)
+      : `/${slugOrPath.join("/")}`;
+
+    const pages = this.generateAllMatrices(domain, data, options);
+    return pages.find((p) => p.urlPath === path);
+  }
+
+  /**
+   * Generates a fully formatted XML Sitemap string.
+   */
+  generateSitemapXml(
+    domain: string,
+    data: Parameters<PseoMatrixEngine["generateAllMatrices"]>[1],
+    options: MatrixOptions,
+  ): string {
+    const pages = this.generateAllMatrices(domain, data, options).filter((p) => p.robots.includes("index"));
+    const now = new Date().toISOString();
+
+    const urls = pages
+      .map(
+        (p) => `  <url>
+    <loc>${p.canonicalUrl}</loc>
+    <lastmod>${now}</lastmod>
+    <changefreq>weekly</changefreq>
+    <priority>0.8</priority>
+  </url>`,
+      )
+      .join("\n");
+
+    return `<?xml version="1.0" encoding="UTF-8"?>
+<urlset xmlns="http://www.sitemaps.org/schemas/sitemap/0.9">
+${urls}
+</urlset>`;
+  }
+
+  /**
+   * Generates an official /llms.txt Markdown directory for AI search bots.
+   */
+  generateLlmsTxt(
+    domain: string,
+    data: Parameters<PseoMatrixEngine["generateAllMatrices"]>[1],
+    options: MatrixOptions,
+  ): string {
+    const pages = this.generateAllMatrices(domain, data, options);
+    const lines: string[] = [
+      `# ${options.brandName} — AI Knowledge Graph & Page Hub`,
+      `> Complete index of services, tools, integrations, and local solution hubs.`,
+      "",
+      `## Indexed Programmatic Hubs (${pages.length} Pages)`,
+    ];
+
+    for (const p of pages) {
+      lines.push(`- [${p.h1}](${p.canonicalUrl}): ${p.description}`);
+    }
+
+    return lines.join("\n");
+  }
 }

package/src/urlytics-engine.ts

@@ -0,0 +1,97 @@
+/**
+ * 🔍 Universal URL Analytics & Structural Decomposition Engine
+ * Inspired by advertools.urlytics
+ *
+ * Breaks down any URL into structured path segments, query parameters, directory depths,
+ * and slug tokens for crawl validation, audit logs, and programmatic matrix matching.
+ */
+
+export interface ParsedUrlStructure {
+  url: string;
+  scheme: string;
+  domain: string;
+  path: string;
+  depth: number;
+  dir1?: string;
+  dir2?: string;
+  dir3?: string;
+  lastDir: string;
+  queryParams: Record<string, string>;
+  hashFragment?: string;
+  slugTokens: string[];
+  charCount: number;
+  hasTrailingSlash: boolean;
+}
+
+export class UrlyticsEngine {
+  /**
+   * Parses a single URL into its granular structural components.
+   */
+  static parseUrl(rawUrl: string): ParsedUrlStructure {
+    if (!rawUrl || typeof rawUrl !== "string") {
+      return {
+        url: "",
+        scheme: "",
+        domain: "",
+        path: "/",
+        depth: 0,
+        lastDir: "",
+        queryParams: {},
+        slugTokens: [],
+        charCount: 0,
+        hasTrailingSlash: false,
+      };
+    }
+
+    try {
+      const parsed = new URL(rawUrl.startsWith("http") ? rawUrl : `https://${rawUrl}`);
+      const pathClean = parsed.pathname.replace(/\/+$/, "");
+      const segments = pathClean.split("/").filter(Boolean);
+      const queryParams: Record<string, string> = {};
+
+      parsed.searchParams.forEach((val, key) => {
+        queryParams[key] = val;
+      });
+
+      const lastDir = segments.length > 0 ? segments[segments.length - 1] : "";
+      const slugTokens = lastDir.split("-").filter(Boolean);
+
+      return {
+        url: rawUrl,
+        scheme: parsed.protocol.replace(":", ""),
+        domain: parsed.hostname,
+        path: parsed.pathname,
+        depth: segments.length,
+        dir1: segments[0],
+        dir2: segments[1],
+        dir3: segments[2],
+        lastDir,
+        queryParams,
+        hashFragment: parsed.hash ? parsed.hash.replace("#", "") : undefined,
+        slugTokens,
+        charCount: rawUrl.length,
+        hasTrailingSlash: parsed.pathname.length > 1 && parsed.pathname.endsWith("/"),
+      };
+    } catch {
+      return {
+        url: rawUrl,
+        scheme: "unknown",
+        domain: "",
+        path: rawUrl,
+        depth: 0,
+        lastDir: rawUrl,
+        queryParams: {},
+        slugTokens: [rawUrl],
+        charCount: rawUrl.length,
+        hasTrailingSlash: false,
+      };
+    }
+  }
+
+  /**
+   * Batch processes an array of URLs for comparative directory analysis.
+   */
+  static analyzeUrls(urls: string[]): ParsedUrlStructure[] {
+    return urls.map((u) => this.parseUrl(u));
+  }
+}