SNACK Quick Summary
- Cloudflare’s new controls split AI traffic into Search, Agent and Training use cases.
- The goal is to move beyond “allow everything” or “block everything,” so a site can preserve discovery while limiting model-training collection.
- The options are live for existing customers in zone Settings, with BotBase and Verified bot rules also moving toward finer classification.
Snackgirls editor note
Red: “If AI search cannot find you, that hurts. If training crawlers take everything with no return, that hurts too. Separating search, agents and training is the interesting part.”
AIKO: “The key is behavior-based classification. Indexing, real-time user-agent actions and training collection may look like bot traffic, but the risk and value are not the same.”
Kirari: “For small blogs and shops, this is the familiar feeling: ‘please help people find me, but do not use everything however you want.’ That makes the update easy to understand.”
As AI search and answer engines grow, website owners face a harder tradeoff: staying discoverable while deciding whether their content can be used for model training. Cloudflare’s July 1 announcement argues that those uses should not be treated as one single “AI bot” bucket.
The three behavior buckets matter
Cloudflare frames AI-related automation around Search, Agent and Training. Search collects or indexes content for later answers. Agent traffic acts in real time on a user’s behalf. Training takes content to train or fine-tune models.
That distinction matters because the value exchange changes. Search can still send referral value back; training collection may not send traffic back to the original site. A site owner may reasonably want different rules for each behavior.
Small sites get more choices between visibility and protection
Cloudflare says it wants to move beyond a single “Block AI Bots” switch. Small publishers, fan sites, indie media and shops often cannot afford to disappear from search, but they also may not want every crawler to use their work for training.
The company says the new options are live now for existing customers in zone Settings. The real-world effect still depends on Cloudflare adoption, bot classification quality and whether bot operators separate their crawlers honestly.
BotBase and Verified bot rules are also changing
Cloudflare also introduced BotBase for Enterprise Bot Management: a searchable view of known bots and agents, with plans to become a more direct control center later.
Another change is that “Verified” no longer simply means “allowed by default.” Cloudflare says verified bots will be allowed according to their relevant category. In plain terms, honest identification can earn access, but misuse can cost trust.
Sources and checked date: Checked on July 3, 2026 (KST)
Comments
0No login needed. Edit or delete your comment from the same browser.
All comments 0
한국어 · English · 日本語No comments yet. Start the conversation.