Not all AI bots do the same thing. Jazui groups them by use:
- Training: they collect content to train models (for example GPTBot, CCBot or Google-Extended).
- Search: they index for an AI search engine's answers (for example OAI-SearchBot or PerplexityBot).
- User-triggered: they read a page because someone asked the assistant (for example ChatGPT-User or Claude-User).
Blocking training bots is your call; blocking search or user-triggered bots means AI can't cite you.
What you can do with each
Site settings › AI bots has the full list, with each bot's company, use, and requests and rejections over the last 30 days (those that visited you first). For each one:
| Policy | Effect |
|---|---|
| Allow (default) | Gets your rendered pages like any bot. |
| Block | Gets a 403 from Jazui (X-Jazui-Blocked: ai-bot). Only for bots that go through your integration. |
| Block and robots.txt | Also adds Disallow: / for that bot to the robots.txt Jazui generates, so bots that respect robots.txt don't even try. |
There are bulk actions, including Block training bots. Some (such as Google-Extended or Applebot-Extended) don't visit pages: they only exist in robots.txt, so they only offer «Block and robots.txt».
llms.txt and Markdown for AI
To help AI understand you better, Jazui publishes for each verified site an llms.txt (your site's index for language models) and a Markdown version of every blog article, linked from its <head>. The llms.txt description is written in Site settings › Sitemap and robots.txt. How to publish them on your domain is in Sitemap, robots and llms.txt.
Measuring
In Bot traffic you see how many requests AI bots make, and in the Visits from AI tab, how many people reach your website from ChatGPT, Perplexity, Gemini and other assistants' answers. With How AI sees you you check what an AI bot reads from any page.