Content-Signal

The syntax is one line inside a robots.txt group, for example `Content-Signal: search=yes, ai-input=yes, ai-train=no`. `search` covers indexing for a link-based result, `ai-input` covers using the page to build a generated answer, and `ai-train` covers using it as training data. Cloudflare introduced it in September 2025 and applies a default to the managed robots.txt files it serves, so a large number of sites declare something without anyone having chosen it.

It is a statement of intent, not an access control. Google's John Mueller has said publicly that it has no effect on any crawler or LLM. Nothing enforces it, no engine has committed to honoring it, and a crawler that ignores it breaks no rule. A site that wants a crawler kept out uses a Disallow rule, or blocks the user agent at the edge.

It is still worth declaring deliberately, for two reasons. It is the only machine-readable place a site states its own position on AI use, which matters if the question is ever litigated or standardized. And a default nobody chose is a default nobody has read: we found a Cloudflare-managed Content-Signal on our own site that contradicted our stated intent, which is exactly the kind of silent mismatch a readiness audit exists to surface.

Frequently asked questions

Does Content-Signal stop AI training on my content?

No. It declares a preference and nothing enforces it. Google's John Mueller has said it has no effect on any crawler or LLM. Use a Disallow rule or an edge block if you need enforcement.

Why does my robots.txt have a Content-Signal line I did not add?

Most likely your CDN added it. Cloudflare applies a default to the managed robots.txt files it serves, so the declaration can arrive without a change to your repository.

What are the three Content-Signal values?

search, ai-input and ai-train. Each takes yes or no, and each covers a different use: link-based indexing, generating an answer, and training a model.