Skip to content
Content Formats and Agent APIs

Content-Signal Headers: Publish Your Preferences

2 min read

Content-Signal headers communicate preferences for training, search and AI input. They are machine-readable declarations, not access control.

Set the three preferences

  • ai-train: use for model training.
  • search: use for search.
  • ai-input: use as input/context in AI systems.

Review these settings in citelayer before changing them. The documented default denies training and allows search and AI input. Choose according to your content policy.

Verify the response

curl -I https://example.com/your-page/

Inspect the Content-Signal header on a public frontend response. Check both normal HTML and Markdown requests if you use both. A CDN or reverse proxy may alter headers.

Robots.txt directs compliant crawlers; a WAF controls access; a Content-Signal declares usage preferences. Blocking a training token does not automatically block every search or user-initiated agent, and publication of a preference does not guarantee compliance.

Make a deliberate change

  1. Record the current response and desired preference.
  2. Change the setting and clear relevant response caches.
  3. Read the public header again.
  4. Check that search crawlers and legitimate users can still access intended pages.

See cache verification and the Content Signals specification.

Next step

Continue with the quick start checklist, or check the common questions. For a reproducible problem, send the plugin versions, affected URL, expected result and actual response through Support. Do not include passwords, license keys or customer details.

Reviewed 6 October 2026 against citelayer® Free 1.13.5, WooCommerce add-on 1.2.19 and EDD add-on 1.2.1. Features depend on the active components and settings on your site.