# Get started with AI Crawl Control

> Turn AI Crawl Control on in log mode, see which crawlers and agents visit your site and what would be blocked, then pick a policy and enforce it.

1. **Turn it on in log mode.** Nothing is blocked: every decision is counted as "would" so you can review it first.

**Dashboard**

   Open the domain, go to **AI Crawl Control → Overview** and click **Turn on in log mode**. Later changes are under **Settings**.
**CLI**

   ```sh
   openwaf-cloud ctl zones set example.com ai_agents.enabled=true ai_agents.mode=log --reason "AI Crawl Control in log mode"
   ```
**API**

   ```sh
   curl -sS -X PATCH "https://dash.openwaf.dev/v1/orgs/$ORG/zones/example.com/edge" \
     -H "Authorization: Bearer $OPENWAF_TOKEN" \
     -H "X-Change-Reason: AI Crawl Control in log mode" \
     -H "Content-Type: application/json" \
     -d '{"ai_agents": {"enabled": true, "mode": "log"}}'
   ```
**MCP**

   `ai_agents_settings` with `{"zone": "example.com", "patch": {"enabled": true, "mode": "log"}, "reason": "…"}`. Simulate, then confirm with `confirm=true` and `expected_version`.

2. **Wait for traffic** and open the report: requests per [category](https://developers.openwaf.dev/ai-crawl-control/categories/), decisions (enforced and "would"), the crawlers seen, verified and spoofed.

   ```sh
   openwaf-cloud ctl ai report example.com --range 7d
   openwaf-cloud ctl ai report example.com --view crawlers
   ```
3. **Choose a policy.** Start from a template, such as **Allow search engines and AI search, block training**, and adjust single categories or bots. See [Set the policy](https://developers.openwaf.dev/ai-crawl-control/policy/).

   ```sh
   openwaf-cloud ctl zones set example.com ai_agents.template=search_ok_block_training --reason "Block AI training"
   ```
4. **Check the "would" numbers** for a few days: they show what enforce mode would block or challenge.
5. **Enforce.**

   ```sh
   openwaf-cloud ctl zones set example.com ai_agents.mode=enforce ai_agents.spoofed_challenge=enforce --reason "Enforce the AI policy"
   ```


> **Note: Beta**
>
> AI Crawl Control is in beta and free during early access. Reports keep 30 days of data.


Optional next steps: [prepend your preferences to robots.txt](https://developers.openwaf.dev/ai-crawl-control/robots-txt/), turn on the [trap](https://developers.openwaf.dev/ai-crawl-control/trap-and-tarpit/), and run an [agent readiness](https://developers.openwaf.dev/ai-crawl-control/agent-readiness/) scan.