Cloudflare AI Crawl Control vs Promptwatch (2026): free bot controls or crawl-to-revenue data?
Cloudflare AI Crawl Control is free on every Cloudflare plan and lets you see, allow and block AI bots. Promptwatch joins crawler logs to citations and conversions. Why many sites use both.
Cloudflare AI Crawl Control shows which AI bots fetch your pages and lets you decide who gets in. Promptwatch shows what those fetches turned into: citations in AI answers, visitors and conversions. That's the whole difference. The first is a free control panel at the edge. The second is a measurement platform that happens to use crawler logs as one of its inputs, Cloudflare's included.
These aren't really rivals, and for most sites on Cloudflare the right answer is to use both. Still, buyers ask which one to rely on, so here is a fair comparison. Cloudflare's docs and plans pages were read on 20 October 2026. Promptwatch facts come from its published plans and llms.txt.
Side by side
| Cloudflare AI Crawl Control | Promptwatch | |
|---|---|---|
| What it is | A feature of the Cloudflare dashboard (formerly AI Audit) | AI search visibility platform |
| Price | Free on every Cloudflare plan; paid features from Cloudflare Pro at $20 | Explore free; crawler logs from Professional at $245/mo |
| Crawler data | Requests by crawler, operator and host; status codes; paths; data transfer; content types; robots.txt violations | Crawler hits from ChatGPTBot, ClaudeBot, PerplexityBot, GoogleOther, Meta AI; crawl-to-citation path; error tracking |
| Bot identification | User agent on Free; Bot Management detection on paid plans | Not published |
| Allow or block bots | Yes, per crawler, as a WAF rule | Not published |
| Charge bots | Pay per crawl, closed beta | Not published |
| Log sources | Traffic through Cloudflare only | Cloudflare, AWS CloudFront, Fastly, Vercel, Netlify, Akamai, Google Cloud CDN, custom HTTP |
| AI referrals | Referral metrics on paid plans | Visitor analytics with conversions (script or GTM) |
| Prompt tracking | None | Daily, with volumes, difficulty and query fan-outs |
| Citation data | None | Pages, domains, Reddit, YouTube, offsite mentions |
| Data access | CSV export, GraphQL Analytics API | REST API, MCP server, Looker Studio |
How Cloudflare AI Crawl Control works
If your site runs behind Cloudflare, AI Crawl Control is already on. Cloudflare's docs say it works on all plans with zero configuration. The Overview tab gives total and allowed requests, unsuccessful requests, the most common status code and the most popular path, with crawlers grouped by operator: OpenAI, Anthropic, Google, Microsoft, Meta, ByteDance and others. The Metrics tab charts requests and data transfer per crawler, shows status codes from 2xx to 5xx, and compares the content types bots asked for with what you served.
The status codes are the practical win. If GPTBot gets a 403 on your pricing page because a firewall rule caught it, this is where you'll see it.
Control is the other half. Each crawler can be set to Allow or Block, and blocking writes a WAF custom rule you can extend. Paid plans can send a custom 403 or 402 response. Pay per crawl, which charges a crawler per successful request, is in closed beta (the overview page calls it private beta), and you need a signup or an Enterprise account executive to get in.
The limits are clear in the docs. It only sees traffic that passes through Cloudflare. On the Free plan crawlers are identified by user-agent string, so only well-known self-identifying bots show up. Referral metrics need a paid plan. And it doesn't run prompts or report citations. It shows fetches, not whether AI recommends you.
How Promptwatch works
Promptwatch starts from the answers. It tracks your prompts daily across ChatGPT, Gemini, Claude, Perplexity, Grok, Copilot, Google AI Overviews, AI Mode and others by monitoring the real interfaces. Citation analytics then shows which of your pages, and which third-party pages, Reddit threads and YouTube videos, the engines cite.
Agent Analytics is where the two products meet. Promptwatch takes crawler logs from Cloudflare, and also from AWS CloudFront, Fastly, Vercel, Netlify, Akamai, Google Cloud CDN and custom HTTP sources. It ties each crawl to the citation path, so you can see that ClaudeBot fetched a page on Tuesday and whether that page started showing up in Claude's answers afterwards. Errors are tracked too. Visitor analytics, added by script or GTM template, closes the loop with AI-referred sessions and conversions.
Crawler logs start on Professional at $245 a month (25M logs, 150 prompts, 1M visitor events, 15 AEO articles). Business at $579 raises that to 100M logs. Agency plans include crawler logs from Kick-off at $199 a month (10M logs).
Who should pick which
Use Cloudflare alone if you only need access control. Maybe you want to block training crawlers, check that retrieval bots aren't hitting errors, or try pay per crawl. It's free and already running, and nothing in Promptwatch replaces the allow and block switches.
Use Promptwatch alone if you're not on Cloudflare. Its other log integrations cover the major CDNs and hosts, and you get the citation and conversion data that a crawler dashboard can't give you.
Use both if you're on Cloudflare and AI visibility is a goal rather than a worry. Set access policy in Cloudflare, connect the same Cloudflare logs to Promptwatch, and judge the policy by what it earns in citations and traffic.
Why most readers should start with Promptwatch
The question most teams actually have isn't "which bots visited?" It's "are we being recommended, and is it worth anything?" Cloudflare answers the first question well and for free. Only Promptwatch answers the second. A crawl count that goes up tells you very little until you know whether those pages were cited and whether the citations sent buyers. That join between crawl, citation and conversion is what Promptwatch's Agent Analytics and visitor analytics are built to provide.
Cloudflare's own docs also put the useful referral data behind a paid plan, so the "free" view stops short of traffic. Promptwatch's Professional plan covers that, plus prompt tracking and Content Agents that publish fixes to WordPress, Webflow or Framer.
Verdict
Cloudflare AI Crawl Control is a good tool and the obvious first step for any Cloudflare site. Keep it for access control. For measuring and improving AI visibility, start with Promptwatch Professional and feed it the same Cloudflare logs.
FAQ
Is Cloudflare AI Crawl Control free?
Yes. Cloudflare's docs say it's available on all plans, including Free, without configuration. Referral metrics and custom block responses need a paid plan, starting with Pro at $20.
Can Promptwatch block AI crawlers?
Blocking isn't among Promptwatch's listed features. Use Cloudflare or your CDN to set access, and Promptwatch to see what the crawling produced.
Does Promptwatch work with Cloudflare logs?
Yes. Cloudflare is one of eight log source types Agent Analytics accepts. Plans and setup are on promptwatch.com.