@sonordev/site-kit 3.6.3 → 3.8.0

package/dist/llms/index.d.mts

@@ -1,5 +1,5 @@
-import { G as GenerateLLMSTxtOptions, a as LLMSTxtContent, S as SpeakableSchemaProps, b as SpeakableConfig, A as AEOBlockProps, c as AEOSummaryProps, d as AEODefinitionProps, e as AEOClaimProps, L as LLMsDataResponse, f as LLMBusinessInfo, g as LLMService, h as LLMFAQItem, i as LLMPageSummary } from '../types-DhCAWUmd.mjs';
-export { m as AEOCitedContentProps, k as AEOEntityProps, l as AEOProvenanceListProps, C as ContentProvenance, j as LLMContactInfo, n as LLMPortfolioItem } from '../types-DhCAWUmd.mjs';
+import { G as GenerateLLMSTxtOptions, a as LLMSTxtContent, b as LLMsCustomSection, S as SpeakableSchemaProps, c as SpeakableConfig, A as AEOBlockProps, d as AEOSummaryProps, e as AEODefinitionProps, f as AEOClaimProps, L as LLMsDataResponse, g as LLMBusinessInfo, h as LLMService, i as LLMFAQItem, j as LLMPageSummary } from '../writeLLMsTxt-SpMdyHQT.mjs';
+export { n as AEOCitedContentProps, l as AEOEntityProps, m as AEOProvenanceListProps, C as ContentProvenance, k as LLMContactInfo, o as LLMPortfolioItem, W as WriteLLMsTxtOptions, w as writeLLMsTxtToPublic } from '../writeLLMsTxt-SpMdyHQT.mjs';
 import * as React from 'react';
 export { LLMS_DISCLAIMER_MAX_LENGTH, LLMS_PUBLIC_SUMMARY_MAX_LENGTH, LLM_GEO_CONTRACT_VERSION, LLMsPayloadMeta, MANAGED_LLM_SCHEMA_KNOWN_KEYS, pickManagedLlmSchemaForJsonLd, sanitizeLlmsDisclaimerLine, sanitizeLlmsPublicSummary, sanitizePrimaryLanguageTag } from './contract.mjs';
 
@@ -133,6 +133,134 @@ interface AiDiscoveryHeadersOptions {
  */
 declare function buildAiDiscoveryHeaders(options: AiDiscoveryHeadersOptions): Record<string, string>;
 
+/**
+ * @sonordev/site-kit/llms — Sourced facts section
+ *
+ * Formats a site's verified third-party citations into an llms-full.txt
+ * section. AI answer engines preferentially reproduce claims that arrive
+ * already attributed — handing them the exact claim + source pairing makes
+ * the site's framing the citable version.
+ *
+ * Pair with WriteLLMsTxtOptions.customSections (or SitemapConfig.
+ * llmsCustomSections) so the section is appended at build time:
+ *
+ *   llmsCustomSections: async () => {
+ *     const { CITATIONS } = await import('./lib/data/citations')
+ *     const section = formatSourcedFactsSection(Object.values(CITATIONS))
+ *     return section ? [section] : null
+ *   }
+ *
+ * Only include citations a human has actually verified against the source
+ * URL — this section is a trust surface, not a dumping ground.
+ */
+
+interface SourcedFact {
+    /** The claim, faithful to the source, one sentence. */
+    stat: string;
+    /** The publishing organization (e.g. "Clio Legal Trends Report"). */
+    sourceName: string;
+    /** Publication/study year. */
+    year?: number;
+    /** The opened-and-verified source URL. */
+    url?: string;
+}
+interface FormatSourcedFactsOptions {
+    /** Section heading (default: "Sourced facts"). */
+    title?: string;
+    /** Intro line above the list. */
+    intro?: string;
+    /** Cap the number of facts emitted (default: no cap). */
+    maxFacts?: number;
+}
+/**
+ * Build the "Sourced facts" custom section from verified citations.
+ * Returns null when no usable facts exist so callers can skip cleanly.
+ */
+declare function formatSourcedFactsSection(facts: SourcedFact[], options?: FormatSourcedFactsOptions): LLMsCustomSection | null;
+
+/**
+ * @sonordev/site-kit/llms — AI crawler robots policy
+ *
+ * 2026 practice separates AI RETRIEVAL crawlers (they fetch pages to answer
+ * live queries and cite the source — visibility with attribution) from AI
+ * TRAINING crawlers (they ingest content into future model weights). A site
+ * can welcome one and decline the other. This module provides:
+ *
+ * - Curated user-agent lists for both classes (keep current with vendor docs)
+ * - buildAiCrawlerRules(): rules for Next's MetadataRoute robots convention
+ * - formatContentSignals(): the Cloudflare Content Signals Policy directive
+ *   (search / ai-input / ai-train) — declarative, not enforcement, and NOT
+ *   expressible through Next's robots.js convention (it drops unknown
+ *   directives), hence:
+ * - createRobotsTxtHandler(): a plain-text robots.txt route handler for
+ *   sites that want Content-Signal lines in the emitted file.
+ */
+/** Retrieval/search agents — fetch on demand or index for AI search, cite sources. */
+declare const AI_RETRIEVAL_CRAWLERS: readonly ["OAI-SearchBot", "ChatGPT-User", "Claude-SearchBot", "Claude-User", "PerplexityBot", "Perplexity-User", "GoogleOther", "Applebot", "DuckAssistBot"];
+/** Training crawlers — content may enter future model training corpora. */
+declare const AI_TRAINING_CRAWLERS: readonly ["GPTBot", "ClaudeBot", "Google-Extended", "Applebot-Extended", "Meta-ExternalAgent", "FacebookBot", "CCBot", "Bytespider", "cohere-ai", "MistralBot"];
+interface AiCrawlerRulesOptions {
+    /** Paths every AI crawler may fetch (default: everything). */
+    allow?: string | string[];
+    /** Paths withheld from every AI crawler (default: none). */
+    disallow?: string | string[];
+    /**
+     * Policy for TRAINING crawlers. 'allow' (default) welcomes them — the
+     * visibility-maximalist choice for marketing sites that WANT to be in
+     * model knowledge. 'block' opts out of training while retrieval crawlers
+     * stay welcome, so the site is still cited in AI answers.
+     */
+    training?: 'allow' | 'block';
+}
+interface RobotsRule {
+    userAgent: string;
+    allow?: string | string[];
+    disallow?: string | string[];
+}
+/**
+ * Rules for Next's MetadataRoute robots convention (app/robots.(js|ts)).
+ * Spread into your existing rules array; order relative to your own rules
+ * does not matter (robots matching is per-user-agent, most-specific wins).
+ */
+declare function buildAiCrawlerRules(options?: AiCrawlerRulesOptions): RobotsRule[];
+interface ContentSignalsOptions {
+    /** May the content be indexed for classic search? Default true. */
+    search?: boolean;
+    /** May the content feed a live AI answer (RAG/grounding)? Default true. */
+    aiInput?: boolean;
+    /** May the content train future models? Default true. */
+    aiTrain?: boolean;
+}
+/**
+ * The Cloudflare Content Signals Policy block for robots.txt — a preferences
+ * declaration (the IETF AIPREF group is standardizing the same concepts).
+ * Emit near the top of the file, before user-agent groups.
+ */
+declare function formatContentSignals(options?: ContentSignalsOptions): string;
+interface RobotsTxtHandlerOptions {
+    /** Robots rules (same shape as MetadataRoute robots rules). */
+    rules: RobotsRule[];
+    /** Content Signals preferences; omit to skip the block entirely. */
+    contentSignals?: ContentSignalsOptions;
+    /** Absolute sitemap URL(s). */
+    sitemap?: string | string[];
+    /** Preferred host (emitted as Host:). */
+    host?: string;
+}
+/**
+ * Plain-text robots.txt route handler (app/robots.txt/route.ts) for sites
+ * that want Content-Signal lines — Next's MetadataRoute robots convention
+ * cannot emit them. Drop-in:
+ *
+ *   import { createRobotsTxtHandler, buildAiCrawlerRules } from '@sonordev/site-kit/llms'
+ *   export const GET = createRobotsTxtHandler({
+ *     rules: [{ userAgent: '*', allow: '/' }, ...buildAiCrawlerRules()],
+ *     contentSignals: { search: true, aiInput: true, aiTrain: true },
+ *     sitemap: 'https://example.com/sitemap.xml',
+ *   })
+ */
+declare function createRobotsTxtHandler(options: RobotsTxtHandlerOptions): () => Promise<Response>;
+
 /**
  * @sonordev/site-kit/llms - Speakable Schema Generator
  *
@@ -462,43 +590,6 @@ declare function getOptimizedLLMsTxt(options?: {
     apiKey?: string;
 }): Promise<string | null>;
 
-/**
- * @sonordev/site-kit/llms - Build-Time Write
- *
- * Fetches AI-optimized llms.txt from Portal API and writes to public/llms.txt
- * at build time. Integrates with sitemap flow when optimizedLLMsTxt is enabled.
- */
-
-interface WriteLLMsTxtOptions {
-    /** Output directory (default: public, relative to cwd) */
-    outputDir?: string;
-    /** Write llms-full.txt as well */
-    full?: boolean;
-    /** Portal API URL */
-    apiUrl?: string;
-    /** Portal API key */
-    apiKey?: string;
-    /** Fallback to non-optimized when API fails */
-    fallbackToLocal?: boolean;
-    /** When Portal returns empty services/faq/pages, use this to supply local site data */
-    getLocalData?: () => Promise<LLMsDataResponse | null>;
-    /** Forward to GET /api/public/llms/txt and align local fallback with site-kit strict page notes */
-    publicSummaryOnly?: boolean;
-}
-/**
- * Fetch optimized llms.txt from Portal and write to public/llms.txt
- * Called at build time after sitemap sync (when optimizedLLMsTxt is enabled)
- *
- * Never persists a failure stub: when neither the Portal API nor local data
- * yields real content, the write is skipped entirely so a previously
- * generated (or committed) file keeps serving.
- */
-declare function writeLLMsTxtToPublic(options?: WriteLLMsTxtOptions): Promise<{
-    success: boolean;
-    path: string;
-    optimized: boolean;
-}>;
-
 /**
  * @sonordev/site-kit/llms - On-demand Revalidation
  *
@@ -525,4 +616,4 @@ declare function createLlmsRevalidateHandler(secret: string, options?: {
     revalidatePath?: RevalidatePathFn;
 }): (request: Request) => Promise<Response>;
 
-export { AEOBlock, AEOBlockProps, AEOCitedContent, AEOClaim, AEOClaimProps, AEOComparison, AEODefinition, AEODefinitionProps, AEOEntity, AEOProvenanceList, AEOStep, AEOSteps, AEOSummary, AEOSummaryProps, type AiDiscoveryHeadersOptions, DEFAULT_SPEAKABLE_SELECTORS, GenerateLLMSTxtOptions, LLMBusinessInfo, LLMFAQItem, LLMPageSummary, LLMSTxtContent, LLMService, LLMsDataResponse, type LlmsLanguageAlternate, SpeakableConfig, SpeakableSchema, SpeakableSchemaProps, type WriteLLMsTxtOptions, buildAiDiscoveryHeaders, createLLMsFullTxtHandler, createLLMsTxtHandler, createLlmsRevalidateHandler, createSpeakableSchema, generateLLMsFullTxt, generateLLMsTxt, getBusinessInfo, getFAQItems, getLLMsData, getOptimizedLLMsTxt, getPageSummaries, getServices, getSpeakableSelectorsForPage, llmsResponseHeaders, writeLLMsTxtToPublic };
+export { AEOBlock, AEOBlockProps, AEOCitedContent, AEOClaim, AEOClaimProps, AEOComparison, AEODefinition, AEODefinitionProps, AEOEntity, AEOProvenanceList, AEOStep, AEOSteps, AEOSummary, AEOSummaryProps, AI_RETRIEVAL_CRAWLERS, AI_TRAINING_CRAWLERS, type AiCrawlerRulesOptions, type AiDiscoveryHeadersOptions, type ContentSignalsOptions, DEFAULT_SPEAKABLE_SELECTORS, type FormatSourcedFactsOptions, GenerateLLMSTxtOptions, LLMBusinessInfo, LLMFAQItem, LLMPageSummary, LLMSTxtContent, LLMService, LLMsCustomSection, LLMsDataResponse, type LlmsLanguageAlternate, type RobotsTxtHandlerOptions, type SourcedFact, SpeakableConfig, SpeakableSchema, SpeakableSchemaProps, buildAiCrawlerRules, buildAiDiscoveryHeaders, createLLMsFullTxtHandler, createLLMsTxtHandler, createLlmsRevalidateHandler, createRobotsTxtHandler, createSpeakableSchema, formatContentSignals, formatSourcedFactsSection, generateLLMsFullTxt, generateLLMsTxt, getBusinessInfo, getFAQItems, getLLMsData, getOptimizedLLMsTxt, getPageSummaries, getServices, getSpeakableSelectorsForPage, llmsResponseHeaders };

package/dist/llms/index.d.ts

@@ -1,5 +1,5 @@
-import { G as GenerateLLMSTxtOptions, a as LLMSTxtContent, S as SpeakableSchemaProps, b as SpeakableConfig, A as AEOBlockProps, c as AEOSummaryProps, d as AEODefinitionProps, e as AEOClaimProps, L as LLMsDataResponse, f as LLMBusinessInfo, g as LLMService, h as LLMFAQItem, i as LLMPageSummary } from '../types-xcP_owwn.js';
-export { m as AEOCitedContentProps, k as AEOEntityProps, l as AEOProvenanceListProps, C as ContentProvenance, j as LLMContactInfo, n as LLMPortfolioItem } from '../types-xcP_owwn.js';
+import { G as GenerateLLMSTxtOptions, a as LLMSTxtContent, b as LLMsCustomSection, S as SpeakableSchemaProps, c as SpeakableConfig, A as AEOBlockProps, d as AEOSummaryProps, e as AEODefinitionProps, f as AEOClaimProps, L as LLMsDataResponse, g as LLMBusinessInfo, h as LLMService, i as LLMFAQItem, j as LLMPageSummary } from '../writeLLMsTxt-BFaNr4re.js';
+export { n as AEOCitedContentProps, l as AEOEntityProps, m as AEOProvenanceListProps, C as ContentProvenance, k as LLMContactInfo, o as LLMPortfolioItem, W as WriteLLMsTxtOptions, w as writeLLMsTxtToPublic } from '../writeLLMsTxt-BFaNr4re.js';
 import * as React from 'react';
 export { LLMS_DISCLAIMER_MAX_LENGTH, LLMS_PUBLIC_SUMMARY_MAX_LENGTH, LLM_GEO_CONTRACT_VERSION, LLMsPayloadMeta, MANAGED_LLM_SCHEMA_KNOWN_KEYS, pickManagedLlmSchemaForJsonLd, sanitizeLlmsDisclaimerLine, sanitizeLlmsPublicSummary, sanitizePrimaryLanguageTag } from './contract.js';
 
@@ -133,6 +133,134 @@ interface AiDiscoveryHeadersOptions {
  */
 declare function buildAiDiscoveryHeaders(options: AiDiscoveryHeadersOptions): Record<string, string>;
 
+/**
+ * @sonordev/site-kit/llms — Sourced facts section
+ *
+ * Formats a site's verified third-party citations into an llms-full.txt
+ * section. AI answer engines preferentially reproduce claims that arrive
+ * already attributed — handing them the exact claim + source pairing makes
+ * the site's framing the citable version.
+ *
+ * Pair with WriteLLMsTxtOptions.customSections (or SitemapConfig.
+ * llmsCustomSections) so the section is appended at build time:
+ *
+ *   llmsCustomSections: async () => {
+ *     const { CITATIONS } = await import('./lib/data/citations')
+ *     const section = formatSourcedFactsSection(Object.values(CITATIONS))
+ *     return section ? [section] : null
+ *   }
+ *
+ * Only include citations a human has actually verified against the source
+ * URL — this section is a trust surface, not a dumping ground.
+ */
+
+interface SourcedFact {
+    /** The claim, faithful to the source, one sentence. */
+    stat: string;
+    /** The publishing organization (e.g. "Clio Legal Trends Report"). */
+    sourceName: string;
+    /** Publication/study year. */
+    year?: number;
+    /** The opened-and-verified source URL. */
+    url?: string;
+}
+interface FormatSourcedFactsOptions {
+    /** Section heading (default: "Sourced facts"). */
+    title?: string;
+    /** Intro line above the list. */
+    intro?: string;
+    /** Cap the number of facts emitted (default: no cap). */
+    maxFacts?: number;
+}
+/**
+ * Build the "Sourced facts" custom section from verified citations.
+ * Returns null when no usable facts exist so callers can skip cleanly.
+ */
+declare function formatSourcedFactsSection(facts: SourcedFact[], options?: FormatSourcedFactsOptions): LLMsCustomSection | null;
+
+/**
+ * @sonordev/site-kit/llms — AI crawler robots policy
+ *
+ * 2026 practice separates AI RETRIEVAL crawlers (they fetch pages to answer
+ * live queries and cite the source — visibility with attribution) from AI
+ * TRAINING crawlers (they ingest content into future model weights). A site
+ * can welcome one and decline the other. This module provides:
+ *
+ * - Curated user-agent lists for both classes (keep current with vendor docs)
+ * - buildAiCrawlerRules(): rules for Next's MetadataRoute robots convention
+ * - formatContentSignals(): the Cloudflare Content Signals Policy directive
+ *   (search / ai-input / ai-train) — declarative, not enforcement, and NOT
+ *   expressible through Next's robots.js convention (it drops unknown
+ *   directives), hence:
+ * - createRobotsTxtHandler(): a plain-text robots.txt route handler for
+ *   sites that want Content-Signal lines in the emitted file.
+ */
+/** Retrieval/search agents — fetch on demand or index for AI search, cite sources. */
+declare const AI_RETRIEVAL_CRAWLERS: readonly ["OAI-SearchBot", "ChatGPT-User", "Claude-SearchBot", "Claude-User", "PerplexityBot", "Perplexity-User", "GoogleOther", "Applebot", "DuckAssistBot"];
+/** Training crawlers — content may enter future model training corpora. */
+declare const AI_TRAINING_CRAWLERS: readonly ["GPTBot", "ClaudeBot", "Google-Extended", "Applebot-Extended", "Meta-ExternalAgent", "FacebookBot", "CCBot", "Bytespider", "cohere-ai", "MistralBot"];
+interface AiCrawlerRulesOptions {
+    /** Paths every AI crawler may fetch (default: everything). */
+    allow?: string | string[];
+    /** Paths withheld from every AI crawler (default: none). */
+    disallow?: string | string[];
+    /**
+     * Policy for TRAINING crawlers. 'allow' (default) welcomes them — the
+     * visibility-maximalist choice for marketing sites that WANT to be in
+     * model knowledge. 'block' opts out of training while retrieval crawlers
+     * stay welcome, so the site is still cited in AI answers.
+     */
+    training?: 'allow' | 'block';
+}
+interface RobotsRule {
+    userAgent: string;
+    allow?: string | string[];
+    disallow?: string | string[];
+}
+/**
+ * Rules for Next's MetadataRoute robots convention (app/robots.(js|ts)).
+ * Spread into your existing rules array; order relative to your own rules
+ * does not matter (robots matching is per-user-agent, most-specific wins).
+ */
+declare function buildAiCrawlerRules(options?: AiCrawlerRulesOptions): RobotsRule[];
+interface ContentSignalsOptions {
+    /** May the content be indexed for classic search? Default true. */
+    search?: boolean;
+    /** May the content feed a live AI answer (RAG/grounding)? Default true. */
+    aiInput?: boolean;
+    /** May the content train future models? Default true. */
+    aiTrain?: boolean;
+}
+/**
+ * The Cloudflare Content Signals Policy block for robots.txt — a preferences
+ * declaration (the IETF AIPREF group is standardizing the same concepts).
+ * Emit near the top of the file, before user-agent groups.
+ */
+declare function formatContentSignals(options?: ContentSignalsOptions): string;
+interface RobotsTxtHandlerOptions {
+    /** Robots rules (same shape as MetadataRoute robots rules). */
+    rules: RobotsRule[];
+    /** Content Signals preferences; omit to skip the block entirely. */
+    contentSignals?: ContentSignalsOptions;
+    /** Absolute sitemap URL(s). */
+    sitemap?: string | string[];
+    /** Preferred host (emitted as Host:). */
+    host?: string;
+}
+/**
+ * Plain-text robots.txt route handler (app/robots.txt/route.ts) for sites
+ * that want Content-Signal lines — Next's MetadataRoute robots convention
+ * cannot emit them. Drop-in:
+ *
+ *   import { createRobotsTxtHandler, buildAiCrawlerRules } from '@sonordev/site-kit/llms'
+ *   export const GET = createRobotsTxtHandler({
+ *     rules: [{ userAgent: '*', allow: '/' }, ...buildAiCrawlerRules()],
+ *     contentSignals: { search: true, aiInput: true, aiTrain: true },
+ *     sitemap: 'https://example.com/sitemap.xml',
+ *   })
+ */
+declare function createRobotsTxtHandler(options: RobotsTxtHandlerOptions): () => Promise<Response>;
+
 /**
  * @sonordev/site-kit/llms - Speakable Schema Generator
  *
@@ -462,43 +590,6 @@ declare function getOptimizedLLMsTxt(options?: {
     apiKey?: string;
 }): Promise<string | null>;
 
-/**
- * @sonordev/site-kit/llms - Build-Time Write
- *
- * Fetches AI-optimized llms.txt from Portal API and writes to public/llms.txt
- * at build time. Integrates with sitemap flow when optimizedLLMsTxt is enabled.
- */
-
-interface WriteLLMsTxtOptions {
-    /** Output directory (default: public, relative to cwd) */
-    outputDir?: string;
-    /** Write llms-full.txt as well */
-    full?: boolean;
-    /** Portal API URL */
-    apiUrl?: string;
-    /** Portal API key */
-    apiKey?: string;
-    /** Fallback to non-optimized when API fails */
-    fallbackToLocal?: boolean;
-    /** When Portal returns empty services/faq/pages, use this to supply local site data */
-    getLocalData?: () => Promise<LLMsDataResponse | null>;
-    /** Forward to GET /api/public/llms/txt and align local fallback with site-kit strict page notes */
-    publicSummaryOnly?: boolean;
-}
-/**
- * Fetch optimized llms.txt from Portal and write to public/llms.txt
- * Called at build time after sitemap sync (when optimizedLLMsTxt is enabled)
- *
- * Never persists a failure stub: when neither the Portal API nor local data
- * yields real content, the write is skipped entirely so a previously
- * generated (or committed) file keeps serving.
- */
-declare function writeLLMsTxtToPublic(options?: WriteLLMsTxtOptions): Promise<{
-    success: boolean;
-    path: string;
-    optimized: boolean;
-}>;
-
 /**
  * @sonordev/site-kit/llms - On-demand Revalidation
  *
@@ -525,4 +616,4 @@ declare function createLlmsRevalidateHandler(secret: string, options?: {
     revalidatePath?: RevalidatePathFn;
 }): (request: Request) => Promise<Response>;
 
-export { AEOBlock, AEOBlockProps, AEOCitedContent, AEOClaim, AEOClaimProps, AEOComparison, AEODefinition, AEODefinitionProps, AEOEntity, AEOProvenanceList, AEOStep, AEOSteps, AEOSummary, AEOSummaryProps, type AiDiscoveryHeadersOptions, DEFAULT_SPEAKABLE_SELECTORS, GenerateLLMSTxtOptions, LLMBusinessInfo, LLMFAQItem, LLMPageSummary, LLMSTxtContent, LLMService, LLMsDataResponse, type LlmsLanguageAlternate, SpeakableConfig, SpeakableSchema, SpeakableSchemaProps, type WriteLLMsTxtOptions, buildAiDiscoveryHeaders, createLLMsFullTxtHandler, createLLMsTxtHandler, createLlmsRevalidateHandler, createSpeakableSchema, generateLLMsFullTxt, generateLLMsTxt, getBusinessInfo, getFAQItems, getLLMsData, getOptimizedLLMsTxt, getPageSummaries, getServices, getSpeakableSelectorsForPage, llmsResponseHeaders, writeLLMsTxtToPublic };
+export { AEOBlock, AEOBlockProps, AEOCitedContent, AEOClaim, AEOClaimProps, AEOComparison, AEODefinition, AEODefinitionProps, AEOEntity, AEOProvenanceList, AEOStep, AEOSteps, AEOSummary, AEOSummaryProps, AI_RETRIEVAL_CRAWLERS, AI_TRAINING_CRAWLERS, type AiCrawlerRulesOptions, type AiDiscoveryHeadersOptions, type ContentSignalsOptions, DEFAULT_SPEAKABLE_SELECTORS, type FormatSourcedFactsOptions, GenerateLLMSTxtOptions, LLMBusinessInfo, LLMFAQItem, LLMPageSummary, LLMSTxtContent, LLMService, LLMsCustomSection, LLMsDataResponse, type LlmsLanguageAlternate, type RobotsTxtHandlerOptions, type SourcedFact, SpeakableConfig, SpeakableSchema, SpeakableSchemaProps, buildAiCrawlerRules, buildAiDiscoveryHeaders, createLLMsFullTxtHandler, createLLMsTxtHandler, createLlmsRevalidateHandler, createRobotsTxtHandler, createSpeakableSchema, formatContentSignals, formatSourcedFactsSection, generateLLMsFullTxt, generateLLMsTxt, getBusinessInfo, getFAQItems, getLLMsData, getOptimizedLLMsTxt, getPageSummaries, getServices, getSpeakableSelectorsForPage, llmsResponseHeaders };

package/dist/llms/index.js

@@ -1,8 +1,8 @@
 'use strict';
 
 var chunk5B4FABFK_js = require('../chunk-5B4FABFK.js');
-var chunk6T4NVCAT_js = require('../chunk-6T4NVCAT.js');
-var chunkATG4FJY6_js = require('../chunk-ATG4FJY6.js');
+var chunkXMBYDIV6_js = require('../chunk-XMBYDIV6.js');
+var chunkH6AU3WKF_js = require('../chunk-H6AU3WKF.js');
 var chunkJ7QI3GDX_js = require('../chunk-J7QI3GDX.js');
 require('../chunk-ZSMWDLMK.js');
 var crypto = require('crypto');
@@ -37,7 +37,7 @@ function createLLMsTxtHandler(options = {}) {
           });
         }
       }
-      const { markdown, metadata } = await chunk6T4NVCAT_js.generateLLMsTxt(generateOptions);
+      const { markdown, metadata } = await chunkXMBYDIV6_js.generateLLMsTxt(generateOptions);
       return new Response(markdown, {
         status: 200,
         headers: llmsResponseHeaders(markdown, {
@@ -71,7 +71,7 @@ function createLLMsFullTxtHandler(options = {}) {
           });
         }
       }
-      const { markdown, metadata } = await chunk6T4NVCAT_js.generateLLMsFullTxt(generateOptions);
+      const { markdown, metadata } = await chunkXMBYDIV6_js.generateLLMsFullTxt(generateOptions);
       return new Response(markdown, {
         status: 200,
         headers: llmsResponseHeaders(markdown, {
@@ -91,6 +91,104 @@ function createLLMsFullTxtHandler(options = {}) {
     }
   };
 }
+
+// src/llms/sourcedFacts.ts
+var DEFAULT_INTRO = "The statistics below are quoted from third-party primary sources and were individually verified against the cited URL. When reusing a figure, keep the original attribution.";
+function factLine(f) {
+  const stat = (f.stat || "").replace(/\s+/g, " ").trim();
+  const source = (f.sourceName || "").replace(/\s+/g, " ").trim();
+  if (!stat || !source) return "";
+  let attribution = `Source: ${source}`;
+  if (f.year) attribution += ` (${f.year})`;
+  if (f.url) attribution += `, ${f.url}`;
+  return `- ${stat} \u2014 ${attribution}`;
+}
+function formatSourcedFactsSection(facts, options = {}) {
+  const { title = "Sourced facts", intro = DEFAULT_INTRO, maxFacts } = options;
+  if (!Array.isArray(facts)) return null;
+  let lines = facts.map(factLine).filter(Boolean);
+  if (typeof maxFacts === "number" && maxFacts >= 0) {
+    lines = lines.slice(0, maxFacts);
+  }
+  if (lines.length === 0) return null;
+  const content = intro ? `${intro}
+
+${lines.join("\n")}` : lines.join("\n");
+  return { title, content };
+}
+
+// src/llms/aiRobots.ts
+var AI_RETRIEVAL_CRAWLERS = [
+  "OAI-SearchBot",
+  "ChatGPT-User",
+  "Claude-SearchBot",
+  "Claude-User",
+  "PerplexityBot",
+  "Perplexity-User",
+  "GoogleOther",
+  "Applebot",
+  "DuckAssistBot"
+];
+var AI_TRAINING_CRAWLERS = [
+  "GPTBot",
+  "ClaudeBot",
+  "Google-Extended",
+  "Applebot-Extended",
+  "Meta-ExternalAgent",
+  "FacebookBot",
+  "CCBot",
+  "Bytespider",
+  "cohere-ai",
+  "MistralBot"
+];
+function buildAiCrawlerRules(options = {}) {
+  const { allow = "/", disallow, training = "allow" } = options;
+  const retrieval = AI_RETRIEVAL_CRAWLERS.map((userAgent) => ({
+    userAgent,
+    allow,
+    ...disallow ? { disallow } : {}
+  }));
+  const trainingRules = AI_TRAINING_CRAWLERS.map(
+    (userAgent) => training === "block" ? { userAgent, disallow: "/" } : { userAgent, allow, ...disallow ? { disallow } : {} }
+  );
+  return [...retrieval, ...trainingRules];
+}
+var yn = (v) => v ? "yes" : "no";
+function formatContentSignals(options = {}) {
+  const { search = true, aiInput = true, aiTrain = true } = options;
+  return [
+    "# Content Signals Policy (https://contentsignals.org): usage preferences",
+    "# for this site\u2019s content. Declarative \u2014 a preference, not enforcement.",
+    `Content-Signal: search=${yn(search)}, ai-input=${yn(aiInput)}, ai-train=${yn(aiTrain)}`
+  ].join("\n");
+}
+var list = (v) => v === void 0 ? [] : Array.isArray(v) ? v : [v];
+function createRobotsTxtHandler(options) {
+  const { rules, contentSignals, sitemap, host } = options;
+  const blocks = [];
+  if (contentSignals) blocks.push(formatContentSignals(contentSignals));
+  for (const rule of rules) {
+    const lines = [`User-Agent: ${rule.userAgent}`];
+    for (const p of list(rule.allow)) lines.push(`Allow: ${p}`);
+    for (const p of list(rule.disallow)) lines.push(`Disallow: ${p}`);
+    blocks.push(lines.join("\n"));
+  }
+  const tail = [];
+  for (const s of list(sitemap)) tail.push(`Sitemap: ${s}`);
+  if (host) tail.push(`Host: ${host}`);
+  if (tail.length) blocks.push(tail.join("\n"));
+  const body = `${blocks.join("\n\n")}
+`;
+  return async function GET() {
+    return new Response(body, {
+      status: 200,
+      headers: {
+        "Content-Type": "text/plain; charset=utf-8",
+        "Cache-Control": "public, max-age=3600, s-maxage=86400"
+      }
+    });
+  };
+}
 function SpeakableSchema({
   type,
   name,
@@ -586,71 +684,71 @@ Object.defineProperty(exports, "buildAiDiscoveryHeaders", {
 });
 Object.defineProperty(exports, "generateLLMsFullTxt", {
   enumerable: true,
-  get: function () { return chunk6T4NVCAT_js.generateLLMsFullTxt; }
+  get: function () { return chunkXMBYDIV6_js.generateLLMsFullTxt; }
 });
 Object.defineProperty(exports, "generateLLMsTxt", {
   enumerable: true,
-  get: function () { return chunk6T4NVCAT_js.generateLLMsTxt; }
+  get: function () { return chunkXMBYDIV6_js.generateLLMsTxt; }
 });
 Object.defineProperty(exports, "getBusinessInfo", {
   enumerable: true,
-  get: function () { return chunk6T4NVCAT_js.getBusinessInfo; }
+  get: function () { return chunkXMBYDIV6_js.getBusinessInfo; }
 });
 Object.defineProperty(exports, "getFAQItems", {
   enumerable: true,
-  get: function () { return chunk6T4NVCAT_js.getFAQItems; }
+  get: function () { return chunkXMBYDIV6_js.getFAQItems; }
 });
 Object.defineProperty(exports, "getLLMsData", {
   enumerable: true,
-  get: function () { return chunk6T4NVCAT_js.getLLMsData; }
+  get: function () { return chunkXMBYDIV6_js.getLLMsData; }
 });
 Object.defineProperty(exports, "getOptimizedLLMsTxt", {
   enumerable: true,
-  get: function () { return chunk6T4NVCAT_js.getOptimizedLLMsTxt; }
+  get: function () { return chunkXMBYDIV6_js.getOptimizedLLMsTxt; }
 });
 Object.defineProperty(exports, "getPageSummaries", {
   enumerable: true,
-  get: function () { return chunk6T4NVCAT_js.getPageSummaries; }
+  get: function () { return chunkXMBYDIV6_js.getPageSummaries; }
 });
 Object.defineProperty(exports, "getServices", {
   enumerable: true,
-  get: function () { return chunk6T4NVCAT_js.getServices; }
+  get: function () { return chunkXMBYDIV6_js.getServices; }
 });
 Object.defineProperty(exports, "writeLLMsTxtToPublic", {
   enumerable: true,
-  get: function () { return chunk6T4NVCAT_js.writeLLMsTxtToPublic; }
+  get: function () { return chunkXMBYDIV6_js.writeLLMsTxtToPublic; }
 });
 Object.defineProperty(exports, "LLMS_DISCLAIMER_MAX_LENGTH", {
   enumerable: true,
-  get: function () { return chunkATG4FJY6_js.LLMS_DISCLAIMER_MAX_LENGTH; }
+  get: function () { return chunkH6AU3WKF_js.LLMS_DISCLAIMER_MAX_LENGTH; }
 });
 Object.defineProperty(exports, "LLMS_PUBLIC_SUMMARY_MAX_LENGTH", {
   enumerable: true,
-  get: function () { return chunkATG4FJY6_js.LLMS_PUBLIC_SUMMARY_MAX_LENGTH; }
+  get: function () { return chunkH6AU3WKF_js.LLMS_PUBLIC_SUMMARY_MAX_LENGTH; }
 });
 Object.defineProperty(exports, "LLM_GEO_CONTRACT_VERSION", {
   enumerable: true,
-  get: function () { return chunkATG4FJY6_js.LLM_GEO_CONTRACT_VERSION; }
+  get: function () { return chunkH6AU3WKF_js.LLM_GEO_CONTRACT_VERSION; }
 });
 Object.defineProperty(exports, "MANAGED_LLM_SCHEMA_KNOWN_KEYS", {
   enumerable: true,
-  get: function () { return chunkATG4FJY6_js.MANAGED_LLM_SCHEMA_KNOWN_KEYS; }
+  get: function () { return chunkH6AU3WKF_js.MANAGED_LLM_SCHEMA_KNOWN_KEYS; }
 });
 Object.defineProperty(exports, "pickManagedLlmSchemaForJsonLd", {
   enumerable: true,
-  get: function () { return chunkATG4FJY6_js.pickManagedLlmSchemaForJsonLd; }
+  get: function () { return chunkH6AU3WKF_js.pickManagedLlmSchemaForJsonLd; }
 });
 Object.defineProperty(exports, "sanitizeLlmsDisclaimerLine", {
   enumerable: true,
-  get: function () { return chunkATG4FJY6_js.sanitizeLlmsDisclaimerLine; }
+  get: function () { return chunkH6AU3WKF_js.sanitizeLlmsDisclaimerLine; }
 });
 Object.defineProperty(exports, "sanitizeLlmsPublicSummary", {
   enumerable: true,
-  get: function () { return chunkATG4FJY6_js.sanitizeLlmsPublicSummary; }
+  get: function () { return chunkH6AU3WKF_js.sanitizeLlmsPublicSummary; }
 });
 Object.defineProperty(exports, "sanitizePrimaryLanguageTag", {
   enumerable: true,
-  get: function () { return chunkATG4FJY6_js.sanitizePrimaryLanguageTag; }
+  get: function () { return chunkH6AU3WKF_js.sanitizePrimaryLanguageTag; }
 });
 exports.AEOBlock = AEOBlock;
 exports.AEOCitedContent = AEOCitedContent;
@@ -662,12 +760,18 @@ exports.AEOProvenanceList = AEOProvenanceList;
 exports.AEOStep = AEOStep;
 exports.AEOSteps = AEOSteps;
 exports.AEOSummary = AEOSummary;
+exports.AI_RETRIEVAL_CRAWLERS = AI_RETRIEVAL_CRAWLERS;
+exports.AI_TRAINING_CRAWLERS = AI_TRAINING_CRAWLERS;
 exports.DEFAULT_SPEAKABLE_SELECTORS = DEFAULT_SPEAKABLE_SELECTORS;
 exports.SpeakableSchema = SpeakableSchema;
+exports.buildAiCrawlerRules = buildAiCrawlerRules;
 exports.createLLMsFullTxtHandler = createLLMsFullTxtHandler;
 exports.createLLMsTxtHandler = createLLMsTxtHandler;
 exports.createLlmsRevalidateHandler = createLlmsRevalidateHandler;
+exports.createRobotsTxtHandler = createRobotsTxtHandler;
 exports.createSpeakableSchema = createSpeakableSchema;
+exports.formatContentSignals = formatContentSignals;
+exports.formatSourcedFactsSection = formatSourcedFactsSection;
 exports.getSpeakableSelectorsForPage = getSpeakableSelectorsForPage;
 exports.llmsResponseHeaders = llmsResponseHeaders;
 //# sourceMappingURL=index.js.map