robots.txt keeps AI search out

Whether robots.txt keeps out the bots that answer — ChatGPT, Claude and Perplexity search — or Googlebot and Bingbot. Training bots are shown, not counted. An assistant cannot cite a page it is not allowed to read; blocking its search bot takes the site out of its answers altogether.

WarningAI-005 · AI readiness · checked once for the whole site

Where it usually comes from

A rule meant to stop AI training written too broadly, a plugin's “block AI” switch, or a leftover Disallow: / for one bot.

How to fix it

Allow the search and user bots (OAI-SearchBot, ChatGPT-User, Claude-SearchBot, Claude-User, PerplexityBot) on the pages you want cited.
Keep training bots out separately if you want to (GPTBot, ClaudeBot, Google-Extended): they have their own names.
Check the result on the site’s page in Trakis, under AI access.

How to check it yourself

Open /robots.txt and look for a User-agent line naming the bot — or a User-agent: * group — with Disallow rules that cover your pages. A CDN can also block bots (Cloudflare's AI Crawl Control): check its dashboard too.

What it looks like in a report

Warning robots.txt keeps OAI-SearchBot (ChatGPT search) out of the whole site.

A blanket block written for AI training also caught the search bots; ChatGPT can no longer read or cite the site. (An example shop; the numbers are illustrative.)

How the audit counts it

AI-005 is a warning in the AI readiness category, and weighs moderately in that category’s score. It is a fact about the site as a whole, so it is reported once, not once per page.

If the finding is deliberate on your site, it can be accepted as intentional in the report. It then leaves the list and the score, and comes back on its own if what it shows changes.

Find it on your site

A Trakis audit runs this check and the other 65 on your site, and tells you on how many pages each one sits. Or go back to all the rules.