Guide · Updated October 2026
OAI-AdsBot: The Complete Guide for ChatGPT Ads Advertisers
Everything OpenAI has documented about its ad crawler — plus what it means for your landing page, and how to check yours in 10 seconds.
What is OAI-AdsBot?
OAI-AdsBot is the web crawler OpenAI sends to every landing page submitted through ChatGPT Ads. When you create a campaign in the Ads Manager, OpenAI visits your page to verify three things before your ad can serve:
- Safety & policy compliance — does the page violate OpenAI's advertising policies?
- Content relevance — does the page match your ad copy and context hints?
- Quality signals — is the page real, reachable content (not an error page or an empty shell)?
It announces itself with this user-agent string:
Mozilla/5.0 AppleWebKit/537.36 (KHTML, like Gecko); compatible; OAI-AdsBot/1.0; +https://openai.com/adsbot
OpenAI formally added OAI-AdsBot to its public crawler documentation in April 2026, as ChatGPT Ads rolled out from its February 2026 US launch to 40+ countries. Since May 2026 the platform is fully self-serve with no minimum spend — which means thousands of first-time advertisers are now meeting this crawler, often without knowing it exists.
How OAI-AdsBot differs from GPTBot and OAI-SearchBot
OpenAI operates several crawlers, and conflating them causes real mistakes:
| Crawler | Visits | Purpose | Data used for training? |
|---|---|---|---|
| GPTBot | The open web | LLM training | Yes |
| OAI-SearchBot | The open web | Search answers & citations | No |
| OAI-AdsBot | Only pages submitted as ads | Ad review & relevance | No |
Key takeaway: OAI-AdsBot only visits pages you explicitly submit. If your ad is stuck in review or was silently rejected, this crawler's experience of your page is the first thing to investigate.
Why your page may be invisible to OAI-AdsBot
The most common failure mode isn't a policy violation — it's that the crawler never saw your real page. Three culprits account for most cases:
1. Your WAF or CDN blocks it
Cloudflare, Akamai, DataDome, and Imperva all classify unknown crawlers as bot traffic. Because OpenAI publishes no IP range list for OAI-AdsBot (unlike its other crawlers), bot-protection defaults can challenge or block it. OpenAI's Help Center explicitly requires that landing pages not block OAI-AdsBot — a blocked crawl generally means a failed review.
2. Your page is a JavaScript-only shell
If your site is a client-side rendered SPA, a crawler that doesn't execute JavaScript sees an empty skeleton — no headline, no copy, nothing to score for relevance. Treat OAI-AdsBot as if it does not run JS: your key content (title, meta description, H1, body copy) must be present in the server-delivered HTML.
3. robots.txt disallows it
OpenAI's guidance for advertisers is to allow both OAI-AdsBot and OAI-SearchBot in robots.txt. A blanket Disallow: / or an overly broad bot rule can take your own landing page out of the running.
Pre-submission checklist
- ✓ Landing page returns HTTP 200 to OAI-AdsBot's user-agent (no challenge page, no 403)
- ✓ Title, meta description, H1, and main copy are in the raw HTML
- ✓ robots.txt does not disallow OAI-AdsBot or OAI-SearchBot
- ✓ Page content genuinely matches your ad title, body, and context hints
- ✓ No prohibited-category content (see our rejection reasons guide)
- ✓ Redirect chain is short and the final URL is what you submitted
How to test all of this in 10 seconds
The fastest way is to simulate the crawl. AdBot Inspector fetches your page with the exact OAI-AdsBot user-agent, alongside a normal browser view and a robots.txt check, then reports whether your page is visible, what a policy scan finds, and how relevant your page is to your ad copy — before you spend a dollar on review cycles.
Check your landing page against the real OAI-AdsBot user-agent
Run a free inspection →No signup. 10 requests per hour.
FAQ
Does OAI-AdsBot crawl my whole site?
No. It only visits URLs submitted as ad landing pages, on demand.
Is my content used to train OpenAI models?
No. OpenAI states OAI-AdsBot data is not used for model training.
How do I verify a request really is OAI-AdsBot?
Reverse DNS: the hostname should end in openai.com. There is no public IP list.
My page is fine — why was my ad still rejected?
Crawlability is necessary but not sufficient. See our 7 rejection reasons guide for policy and relevance causes.