Content-Signal is a Cloudflare-originated, emerging robots.txt policy directive. Its core signals are search, ai-input, and ai-train; omitted signals express no preference. An additional use field for retention or reuse is still an evolving extension. These signals state policy preferences and require crawler cooperation or separate enforcement.
In plain terms
search covers building a search index and returning links or short excerpts; it does not include AI-generated summaries.
ai-input covers using content as real-time input for RAG, grounding, or generated search answers; ai-train covers training or fine-tuning models.
A Content-Signal line does not itself block network access. Access control still comes from Allow/Disallow rules and, when needed, edge or application enforcement.
Why it matters
It lets brands invite the uses they want, like citation and grounding, while expressing preferences about others, without shutting crawlers out entirely.
As content-usage norms mature, an explicit signal documents your intent for the agents that respect it.