AI Crawler robots.txt Generator

Block the bots that train on your content. Keep the ones that send you traffic. Tick what you want blocked and copy the result — everything runs in your browser, nothing is uploaded.

The mistake almost everyone makes: OpenAI, Anthropic and Google each run separate bots for training and for search. Blocking OAI-SearchBot or Claude-SearchBot does not stop training — it just deletes you from ChatGPT and Claude search results. The presets below keep the search bots and block the training crawlers.

1. Choose what to block

2. Copy your robots.txt

Save this as robots.txt at the root of your domain (https://yoursite.com/robots.txt).

robots.txt is a request, not a fence

Well-behaved crawlers honour it. Scrapers ignore it entirely, and a user-agent string takes one line of code to fake. If you need the block enforced, you need a rule at your web server or CDN, and then you need to verify it actually fired in production.

That is what the paid kit is for:

AI Crawler Control Kit — $9

  • The full annotated table below, with what each bot is really for and the cost of blocking it
  • Ready-to-paste enforcement rules for nginx, Apache, Caddy and Cloudflare WAF
  • verify-blocks.sh — curls your site as each bot and tells you PASS/FAIL, so you can prove the block works and that you did not accidentally 403 Googlebot
  • Log-monitoring one-liners: which AI bots are hitting you now, how much bandwidth each is burning, and which paths they hammer
  • An llms.txt template for the opt-in side, with the licensing line most people omit

Get the kit — $9

Delivered by email within 24 hours of purchase — reply to your Stripe receipt, or write to lovihs1z1t (at) emalupe.com with your receipt if you would like it sooner. Not what you expected? Ask and you get a refund, no argument.

Every AI crawler that matters in 2026

Free reference. keep means blocking it costs you traffic or citations; block is a training or resale crawler; token is a robots.txt-only opt-out that never appears as a real user-agent — so blocking it in nginx or a WAF does nothing at all.

User-agentRun byWhat it does

Notes that save you a support ticket

More free tools on Smeltworks