cd /news/ai-crawlers/is-cloudflare-blocking-chatgpt-from-… · home › topics › ai-crawlers › article
[ARTICLE · art-173373] src=dev.to ↗ pub= topic=ai-crawlers verified=true sentiment=· neutral

Is Cloudflare Blocking ChatGPT From Your Site? Check in 30 Seconds

A developer audited 286 developer portfolios and found that 19 blocked GPTBot and ClaudeBot training crawlers, with 17 of those blocks written by Cloudflare's managed robots.txt rather than by site owners. The writeup distinguishes training crawlers (GPTBot, ClaudeBot, Google-Extended) from AI search crawlers (OAI-SearchBot, ChatGPT-User, Claude-SearchBot, PerplexityBot) and offers robots.txt inspection plus curl user-agent tests to detect firewall-level blocks such as Cloudflare's AI Crawl Control returning 403s.

key takeaway

Of 286 audited portfolios, 19 blocked GPTBot and ClaudeBot and 17 of those blocks came from Cloudflare's managed robots.txt, so check whether your training-crawler blocks were deliberate and test AI search crawlers with a curl user-agent request to catch firewall-level 403s.

by read3 min views4 publishedNov 20, 2026

Short answer: probably not ChatGPT search, but probably ChatGPT training. Cloudflare's managed robots.txt blocks training crawlers like GPTBot and ClaudeBot. It doesn't block OAI-SearchBot or ChatGPT-User, which fetch pages to answer and cite. Open yourdomain.com/robots.txt to check the first part, and run the curl below to check the second.

When I audited 286 developer portfolios for an earlier study, 19 blocked GPTBot and ClaudeBot. When I opened those files, 17 of the 19 blocks had been written by Cloudflare, not by the owner. Most of those people never decided to block AI. A dashboard setting decided for them.

This is the part most posts get wrong. "AI bots" aren't one thing:

Crawler Company What it does
GPTBot OpenAI Collects content to train models
OAI-SearchBot OpenAI Indexes pages for ChatGPT search
ChatGPT-User OpenAI Fetches a page when a user asks about it
ClaudeBot Anthropic Collects content to train models
Claude-SearchBot Anthropic Indexes pages for Claude search
PerplexityBot Perplexity Indexes pages for Perplexity answers
Google-Extended Controls whether Gemini trains on your content (Google Search isn't affected)

Blocking a training crawler keeps you out of future models. Blocking a search crawler keeps you out of the answers. Those are very different outcomes.

Open https://yourdomain.com/robots.txt in your browser. If Cloudflare manages it, you'll see something like this at the top:

User-Agent: *
Content-signal: search=yes,ai-train=no
Allow: /

User-agent: GPTBot
Disallow: /

User-agent: ClaudeBot
Disallow: /

User-agent: Google-Extended
Disallow: /

That's the training list. Search crawlers aren't in it, so with only this block, ChatGPT search and Perplexity can still read you. The Content-signal line says the same thing in words: search yes, training no.

If you see Disallow: / under OAI-SearchBot, ChatGPT-User or PerplexityBot, that one is blocking you from AI answers. Someone added it by hand or with a plugin.

robots.txt is only a request. Cloudflare also has AI Crawl Control, which blocks crawlers at the firewall with a 403. That doesn't show up in robots.txt at all, so you have to ask as the bot:

curl -s -o /dev/null -w "%{http_code}\n" -A "OAI-SearchBot/1.0; +https://openai.com/searchbot" https://yourdomain.com/
curl -s -o /dev/null -w "%{http_code}\n" -A "Mozilla/5.0" https://yourdomain.com/

If the second line prints 200 and the first prints 403 (or 402), you're blocked at the firewall. Swap in ChatGPT-User, PerplexityBot or GPTBot to test the others. On Windows, use curl.exe and NUL instead of /dev/null.

This isn't perfect. Cloudflare can also verify real bots by IP, so a fake user agent from your laptop won't always get the same answer as the real crawler. But a 403 here is a strong hint to go look at the dashboard.

At the time of writing, both settings live in the Cloudflare dashboard for your domain:

My take for a portfolio or a small business site: allow the search crawlers, because that's how ChatGPT, Claude and Perplexity find and recommend you. Training is a real choice, and blocking it is reasonable. Just make it on purpose.

To allow everything explicitly in a robots.txt you write yourself:

User-agent: OAI-SearchBot
User-agent: ChatGPT-User
User-agent: Claude-SearchBot
User-agent: PerplexityBot
Allow: /

Run your site through the free audit. It reads your robots.txt and tells you which AI search crawlers and which training crawlers are blocked, separately, in about ten seconds and with no sign-up.

── more in #ai-crawlers 4 stories · sorted by recency
── more on @cloudflare 3 stories trending now
sponsored brought to you by zahid.host 4,200+ EU-deployed projects
reading about agents? ship yours in a single git push.

Run your AI side-project on zahid.host

EU-based hosting, git-push deploys, automatic HTTPS, no cold starts. Free tier with a custom domain — perfect for shipping the agent you just read about.

$git push zahid main
→ Live at https://your-agent.zahid.host ✓
Get free account → Pricing
from €0/mo · no card required
LIVE [news/is-cloudflare-blocki…] indexed:0 read:3min 2026-11-20 · —