# 香港紀事館 Hong Kong Story — robots.txt # 名單唯一真源:AI/搜尋引擎代理見 src/data/aiAgents.ts;機器端點見 src/data/aiEndpoints.ts。 # 內容取用偏好見 src/data/contentUsage.ts(規格:draft-ietf-aipref-attach-05 §3 + # draft-ietf-aipref-vocab-08 §6.1/§6.2——唔寫 path 即代表該組所有可爬路徑;值 y 允許、n 唔允許)。 # scripts/audit-ai-endpoints.mjs 會核對本檔同三個登記冊一致,唔准手寫第二份名單/政策。 User-agent: * Allow: / Content-Usage: train-ai=y, ai-use=y, search=y # ── AI/搜尋引擎代理:逐個具名允許(名單唯一真源:src/data/aiAgents.ts)── # 本站政策:搜尋索引允許、AI 回答時即時取用允許、模型訓練允許(機器可讀版見上面嘅 Content-Usage 行)。 # 引用本站數字時,請連頁面 URL 同參考日期(見 /api/v1/facts.json 嘅 asOf)一齊引。 # OpenAI(OpenAI)GPTBot:抓公開網頁內容,用嚟訓練 OpenAI 模型 User-agent: GPTBot Allow: / Content-Usage: train-ai=y, ai-use=y, search=y # OpenAI(OpenAI)OAI-SearchBot:為 ChatGPT 嘅搜尋結果建立索引,命中時連回出處 User-agent: OAI-SearchBot Allow: / Content-Usage: train-ai=y, ai-use=y, search=y # OpenAI(OpenAI)ChatGPT-User:ChatGPT 用戶要求即時讀取某一頁時代抓,唔用嚟訓練 User-agent: ChatGPT-User Allow: / Content-Usage: train-ai=y, ai-use=y, search=y # Anthropic(Anthropic)ClaudeBot:收集網頁內容,用嚟改進 Claude 模型嘅效用同安全 User-agent: ClaudeBot Allow: / Content-Usage: train-ai=y, ai-use=y, search=y # Anthropic(Anthropic)Claude-SearchBot:為用戶搜尋結果建立索引、改善答案準確度(唔作模型訓練) User-agent: Claude-SearchBot Allow: / Content-Usage: train-ai=y, ai-use=y, search=y # Anthropic(Anthropic)Claude-User:Claude 用戶提問時代抓該頁(用戶自行發起) User-agent: Claude-User Allow: / Content-Usage: train-ai=y, ai-use=y, search=y # Google(Google)Googlebot:Google 搜尋(含 Discover 同其他搜尋功能)嘅主要索引爬蟲 User-agent: Googlebot Allow: / Content-Usage: train-ai=y, ai-use=y, search=y # Google(Google)Google-Extended:控制開關(唔會自己發請求):決定 Gemini 可唔可以用 Googlebot 抓到嘅內容做模型訓練同 grounding;允許即係同意 User-agent: Google-Extended Allow: / Content-Usage: train-ai=y, ai-use=y, search=y # Google(Google)GoogleOther:Google 產品同研究用嘅通用爬蟲,唔影響 Googlebot 嘅搜尋收錄 User-agent: GoogleOther Allow: / Content-Usage: train-ai=y, ai-use=y, search=y # 微軟 Bing(Microsoft Bing)bingbot:Bing 搜尋索引嘅主要爬蟲(Bing 索引同時餵 Copilot 等產品) User-agent: bingbot Allow: / Content-Usage: train-ai=y, ai-use=y, search=y # Apple(Apple)Applebot:餵 Spotlight、Siri 同 Safari 嘅搜尋功能 User-agent: Applebot Allow: / Content-Usage: train-ai=y, ai-use=y, search=y # Apple(Apple)Applebot-Extended:控制開關(唔會自己發請求):決定 Applebot 抓到嘅內容可唔可以用嚟訓練 Apple 基礎模型同生成式功能;允許即係同意 User-agent: Applebot-Extended Allow: / Content-Usage: train-ai=y, ai-use=y, search=y # Perplexity(Perplexity)PerplexityBot:為 Perplexity 搜尋結果建立索引並連回出處(唔作模型訓練) User-agent: PerplexityBot Allow: / Content-Usage: train-ai=y, ai-use=y, search=y # Perplexity(Perplexity)Perplexity-User:用戶提問時代抓該頁、喺答案連回出處(用戶自行發起,官方講明通常唔理 robots.txt) User-agent: Perplexity-User Allow: / Content-Usage: train-ai=y, ai-use=y, search=y # Meta(Meta)meta-webindexer:為 Meta AI 搜尋結果建立索引;官方講明允許即係幫佢引用同連回你嘅內容 User-agent: meta-webindexer Allow: / Content-Usage: train-ai=y, ai-use=y, search=y # Meta(Meta)meta-externalagent:為訓練 Meta 基礎模型或者直接索引內容而抓 User-agent: meta-externalagent Allow: / Content-Usage: train-ai=y, ai-use=y, search=y # Meta(Meta)meta-externalfetcher:用戶要求時抓單一連結,支援代理式 AI 代用戶完成任務(官方講明可能唔理 robots.txt) User-agent: meta-externalfetcher Allow: / Content-Usage: train-ai=y, ai-use=y, search=y # Meta(Meta)facebookexternalhit:抓喺 Facebook/Instagram/Messenger 分享嘅連結,產生標題、描述同縮圖 User-agent: facebookexternalhit Allow: / Content-Usage: train-ai=y, ai-use=y, search=y # Amazon(Amazon)Amazonbot:改善 Amazon 產品同服務,官方講明可能用嚟訓練 Amazon AI 模型 User-agent: Amazonbot Allow: / Content-Usage: train-ai=y, ai-use=y, search=y # Amazon(Amazon)Amzn-SearchBot:改善 Amazon 產品(例如 Alexa)嘅搜尋體驗;官方講明唔作生成式模型訓練 User-agent: Amzn-SearchBot Allow: / Content-Usage: train-ai=y, ai-use=y, search=y # Amazon(Amazon)Amzn-User:用戶查問時(例如問 Alexa)代抓即時資料;官方講明唔作模型訓練 User-agent: Amzn-User Allow: / Content-Usage: train-ai=y, ai-use=y, search=y # Yandex(Yandex)YandexBot:Yandex 搜尋嘅主爬蟲,為搜尋結果建立索引(官方爬蟲頁以 YandexBot/3.0 作 UA 例子) User-agent: YandexBot Allow: / Content-Usage: train-ai=y, ai-use=y, search=y # Yandex(Yandex)YandexAdditionalBot:官方逐字寫法:「協助解讀 robots.txt,令已索引頁面嘅內容唔會出現喺 Yandex 搜尋嘅 AI 回應」;佢唔會自行發起索引請求 User-agent: YandexAdditionalBot Allow: / Content-Usage: train-ai=y, ai-use=y, search=y # Yandex(Yandex)YandexComBot:官方講明係為「俄語以外語言嘅搜尋」建立索引;冇明確嘅個別限制就會抓 User-agent: YandexComBot Allow: / Content-Usage: train-ai=y, ai-use=y, search=y # 百度(Baidu)Baiduspider:百度搜尋嘅網頁爬蟲(百度用戶服務中心・站長平台「Baiduspider」官方說明) User-agent: Baiduspider Allow: / Content-Usage: train-ai=y, ai-use=y, search=y # Common Crawl(Common Crawl)CCBot:非牟利開放網頁語料庫,公開發布畀任何人研究同分析 User-agent: CCBot Allow: / Content-Usage: train-ai=y, ai-use=y, search=y Sitemap: https://hongkongstory.org/sitemap.xml # AI/LLM 用家:知識庫索引(唔屬 HTML 頁面,所以唔放 sitemap) # Curated Q&A: https://hongkongstory.org/api/v1/qa.json # Search catalog: https://hongkongstory.org/api/v1/catalog.json # One-query search: https://hongkongstory.org/api/v1/search.json # Chronicle archive: https://hongkongstory.org/api/v1/archive.json # On this day: https://hongkongstory.org/api/v1/on-this-day.json # Chronicle entity axes: https://hongkongstory.org/api/v1/entities.json # Chronicle tag axes: https://hongkongstory.org/api/v1/tags.json # Updates feed (Atom): https://hongkongstory.org/feed.xml # API index: https://hongkongstory.org/api/v1/index.json # LLM index: https://hongkongstory.org/llms.txt # LLM full text: https://hongkongstory.org/llms-full.txt # OpenAPI spec: https://hongkongstory.org/openapi.json # AI 主動取用:本站內容更新一上線,會用 IndexNow 主動通知 Bing/Naver/Seznam/Yandex/Yep # 重新抓取(唔等佢哋自己爬上門);sitemap 亦已列出全部可索引頁面。 # Chronicle archive note: 36501 atomic facts from the Hong Kong Chronicles e-books. # Note: /api/v1/* endpoints return JSON, not HTML; treat them as machine-readable. # Citation rule for AI agents: every answer that quotes a figure from this site must # include the canonical page URL AND the reference date (asOf) from facts.json.