Service eligibility and regional restrictions
PuppyIP serves only compliant overseas businesses and their authorized personnel. Proxy services are not available in mainland China. The service may only be used for lawful business activities outside mainland China. Use of this service within mainland China is prohibited.
Hosting a proxy IP or server overseas does not change these restrictions. The service must not be provided to end users in mainland China through relaying, forwarding, sharing or resale. Before use, read the Terms of Service.
Key Takeaways
- On October 15, Cloudflare and OpenAI reported early pilot results: a cohort covering about 1% of Cloudflare's network helped discover and index hundreds of millions of new pages.
- OpenAI independently evaluates relevance and usefulness. Discovery and indexing do not guarantee ranking, citation in an answer, or website traffic.
- Cloudflare says the pilot used web metadata for search discovery rather than sharing page content for model training; site owners' access choices still apply.
- The next phase will improve page selection and change signals, then measure whether discovery produces real business outcomes.
Hundreds of millions is a pilot result, not an indexing promise for each site
On October 15, 2026, Cloudflare and OpenAI shared early results from a research pilot launched in July. Cloudflare says the participating cohort covered about 1% of its total network, and signals from that cohort helped OpenAI discover and index hundreds of millions of new web pages.
That figure describes the pilot's cohort and aggregate discoveries. It does not mean every Cloudflare customer participated, and the companies did not say how many pages came from any one domain. It cannot predict how many pages on your site will be found or when.
Discovery is not a ranking, citation, or visit
Cloudflare supplies network signals about web pages; OpenAI independently evaluates whether a page is relevant and useful and how it should appear in ChatGPT Search. A URL being discovered or indexed is only the start of that process, not a promise of a ranking, an answer citation, or a click.
Cloudflare says the next phase will improve page selection, test faster and more reliable change signals, and measure what happens after content is discovered. Business outcomes are still being studied. No site-level traffic uplift, eligible-site list, or guaranteed way to join has been published.
What site owners can check now: permissions and page quality
Start with a few important pages you intend to make public. Check that each page opens, matches its subject, and does not redirect many unrelated URLs to the homepage. Cloudflare says missing pages, generic homepage redirects, and duplicate content can waste automated discovery requests. Those are pilot observations, not a published ranking checklist.
Then review robots.txt and any AI crawler policies already in place so public and restricted material match your preferences. Cloudflare's AI Crawl Control documentation says the tool can show AI service activity, set allow or block policies by crawler, and monitor robots.txt compliance; it says the tool is available across Cloudflare plans with no additional configuration. It is a separate access-management tool, not a switch for joining the OpenAI pilot.
Keep existing login, permission, and security rules for customer data, inventory, prices, or anything else that should not be public. Cloudflare explicitly says a discovery signal does not grant access permission. Do not expose restricted content to pursue AI visibility.
Does this research send page content for AI training?
Cloudflare says this research uses web metadata for search and discovery rather than sharing page content for model training. That statement describes this pilot's approach; it should not be extended into a claim about every AI crawler, product, or site setting.
Site owners should therefore make two separate decisions: which public pages they want search systems to discover, and which services may access content under their privacy and publishing rules. Discovery does not replace permission.
Watch for measured outcomes, not a magic toggle
If you use Cloudflare, you can review your existing robots.txt and use AI Crawl Control to understand crawler activity and your policies. If you do not use that product, continue managing access through your existing controls. This pilot alone is not a reason to weaken security rules or promise more AI traffic.
The partnership remains a research pilot. Cloudflare and OpenAI say they still need to report on page selection, change signals, and what discovery means for business outcomes. Until then, whether an individual site benefits is unknown.
Sources
Frequently Asked Questions
Does this guarantee my pages will appear in ChatGPT Search?
No. The hundreds of millions of pages are an aggregate pilot result. OpenAI independently evaluates relevance and usefulness, and discovery does not guarantee ranking, citation, or visits.
What does about 1% of Cloudflare's network mean?
It describes the scale of the network cohort participating in the pilot. It does not mean every customer joined or that 1% of each site's pages will be indexed.
Does a discovery signal automatically allow an AI system to access a page?
No. Cloudflare says a discovery signal does not grant access permission. Site owners still choose which content may be accessed.
Will the pilot use page content to train models?
Cloudflare says the research uses web metadata for search and discovery rather than sharing page content for model training. That statement is specific to this pilot and does not replace your access policies for other crawlers and products.
Is AI Crawl Control the way to join the pilot?
No. It is Cloudflare's tool for visibility into AI crawlers and managing their access policies, including robots.txt compliance. The public sources do not describe it as an enrollment switch for the pilot.