diff --git a/docs/content/2.guides/7.robot-recipes.md b/docs/content/2.guides/7.robot-recipes.md
index d3e1c193..5731f86f 100644
--- a/docs/content/2.guides/7.robot-recipes.md
+++ b/docs/content/2.guides/7.robot-recipes.md
@@ -49,7 +49,9 @@ export default defineNuxtConfig({
 })
 ```
 
-This will block the following AI crawlers: `GPTBot`, `ChatGPT-User`, `Claude-Web`, `anthropic-ai`, `Applebot-Extended`, `Bytespider`, `CCBot`, `cohere-ai`, `Diffbot`, `FacebookBot`, `Google-Extended`, `ImagesiftBot`, `PerplexityBot`, `OmigiliBot`, `Omigili`
+This will block the AI crawlers listed in the `AiBots` constant, which follows the [ai.robots.txt](https://github.com/ai-robots-txt/ai.robots.txt) reference list: `GPTBot`, `ChatGPT-User`, `OAI-SearchBot`, `ClaudeBot`, `Claude-User`, `Claude-SearchBot`, `Claude-Web`, `anthropic-ai`, `Google-Extended`, `Google-CloudVertexBot`, `Google-NotebookLM`, `Gemini-Deep-Research`, `GoogleOther`, `meta-externalagent`, `meta-externalfetcher`, `FacebookBot`, `Amazonbot`, `bedrockbot`, `Applebot-Extended`, `PerplexityBot`, `Perplexity-User`, `Bytespider`, `TikTokSpider`, `omgili`, `omgilibot`, `Webzio-Extended`, `YandexAdditional`, `YandexAdditionalBot`, `CCBot`, `cohere-ai`, `Diffbot`, `ImagesiftBot`, `MistralAI-User`, `DeepSeekBot`, `AI2Bot`, `Ai2Bot-Dolma`, `YouBot`, `Timpibot`, `PanguBot`, `DuckAssistBot`, `FirecrawlAgent`, `ICC-Crawler`, `Kangaroo Bot`, `SemrushBot-OCOB`, `img2dataset`
+
+Search crawlers are not affected. `Googlebot`, `Applebot` (Siri and Spotlight) and `facebookexternalhit` (Open Graph link previews) are deliberately left out, so blocking AI bots never costs you search visibility or social previews.
 
 ### Blocking Privileged Pages
 
diff --git a/src/const.ts b/src/const.ts
index dd3abbed..91047050 100644
--- a/src/const.ts
+++ b/src/const.ts
@@ -19,20 +19,71 @@ export const NonHelpfulBots = [
   'magpie-crawler',
 ]
 
+/**
+ * A list of AI crawler user agents, both training and retrieval.
+ *
+ * Search crawlers are deliberately excluded, so `blockAiBots` never costs a site its search
+ * visibility. This is why `Googlebot`, `Applebot` (Siri, Spotlight) and `facebookexternalhit`
+ * (Open Graph previews) are absent, while their AI-only counterparts `Google-Extended` and
+ * `Applebot-Extended` are listed.
+ *
+ * @credits https://github.com/ai-robots-txt/ai.robots.txt
+ */
 export const AiBots = [
+  // OpenAI
   'GPTBot',
   'ChatGPT-User',
+  'OAI-SearchBot',
+  // Anthropic
+  'ClaudeBot',
+  'Claude-User',
+  'Claude-SearchBot',
   'Claude-Web',
   'anthropic-ai',
+  // Google
+  'Google-Extended',
+  'Google-CloudVertexBot',
+  'Google-NotebookLM',
+  'Gemini-Deep-Research',
+  'GoogleOther',
+  // Meta
+  'meta-externalagent',
+  'meta-externalfetcher',
+  'FacebookBot',
+  // Amazon
+  'Amazonbot',
+  'bedrockbot',
+  // Apple
   'Applebot-Extended',
+  // Perplexity
+  'PerplexityBot',
+  'Perplexity-User',
+  // ByteDance
   'Bytespider',
+  'TikTokSpider',
+  // Webz.io
+  'omgili',
+  'omgilibot',
+  'Webzio-Extended',
+  // Yandex
+  'YandexAdditional',
+  'YandexAdditionalBot',
+  // Others
   'CCBot',
   'cohere-ai',
   'Diffbot',
-  'FacebookBot',
-  'Google-Extended',
   'ImagesiftBot',
-  'PerplexityBot',
-  'OmigiliBot',
-  'Omigili',
+  'MistralAI-User',
+  'DeepSeekBot',
+  'AI2Bot',
+  'Ai2Bot-Dolma',
+  'YouBot',
+  'Timpibot',
+  'PanguBot',
+  'DuckAssistBot',
+  'FirecrawlAgent',
+  'ICC-Crawler',
+  'Kangaroo Bot',
+  'SemrushBot-OCOB',
+  'img2dataset',
 ]