Will my site's bot protection or AI crawler blocks stop my ChatGPT ads from being approved?

TofuAnswer by the Tofu team · checked

They can. OpenAI checks ChatGPT ad landing pages with a crawler called OAI-AdsBot, and its ad docs say your destination must be "accessible to OpenAI's ad crawlers." A robots.txt rule, a firewall or bot-protection challenge (Cloudflare, Akamai and similar) or a login wall can stop that check. Blocking GPTBot, OpenAI's training crawler, doesn't block OAI-AdsBot: they are separate user agents. To fix it, allow OAI-AdsBot in robots.txt and your firewall, confirm it can load the page, then edit or resubmit the ad so it's reviewed again.

Ask Tofu next

Pick a question and Tofu Ads Agent answers it in chat.

As of November 2026.

Before you start

  • The exact landing page URLs your ads use, including any tracking parameters.
  • Access to robots.txt and your CDN, firewall or bot-protection settings, or the person who runs them. Most fixes happen there.
  • Access to ChatGPT Ads Manager to see the ad's review status and edit it.

Step 1: Know which OpenAI crawler checks ads

OpenAI's crawler page lists four user agents, each separate from the others. Only one exists for ads:

User agentWhat OpenAI uses it forNeeded for ChatGPT ads?
OAI-AdsBotChecks pages submitted as ads against OpenAI's policies, and may use the page's content to decide when to show the ad. Visits only ad landing pages, and its data isn't used to train modelsYes
OAI-SearchBotFinds sites to show in ChatGPT search answersRecommended by OpenAI's advertiser help page (see Step 3)
GPTBotCrawls content that may be used to train OpenAI's modelsNot mentioned in OpenAI's ads guidance
ChatGPT-UserVisits a page when a ChatGPT user asks it toNot mentioned in OpenAI's ads guidance

OAI-AdsBot identifies itself as Mozilla/5.0 AppleWebKit/537.36 (KHTML, like Gecko); compatible; OAI-AdsBot/1.0; +https://openai.com/adsbot. OpenAI publishes its IP ranges at openai.com/adsbot.json.

So yes, you can keep GPTBot blocked and still run ChatGPT ads, as long as the block names GPTBot and nothing else.

Step 2: Check robots.txt

Open yoursite.com/robots.txt and look for two things:

  • A group for OAI-AdsBot that disallows your landing pages.
  • A User-agent: * group with Disallow: /, or a disallow that covers your landing pages. Under the robots.txt standard (RFC 9309), a crawler with no group of its own must follow the * group. So if OAI-AdsBot follows robots.txt, a blanket block catches it even if you never named it.

OpenAI's crawler page doesn't say whether OAI-AdsBot follows robots.txt. Its advertiser help page, in the version we saw in search results, says OpenAI's crawlers respect robots.txt and stop as soon as access is disallowed. OpenAI's help center couldn't be read to confirm this.

To let OAI-AdsBot in while keeping other blocks, give it its own group. A crawler that has a group of its own ignores the * group:

User-agent: OAI-AdsBot
Allow: /

User-agent: GPTBot
Disallow: /

The GPTBot group is optional. It keeps your training block in place.

Step 3: Check your firewall, CDN and bot protection

robots.txt is a request. Firewalls and bot tools block for real, and many block or challenge AI crawlers by default.

OpenAI's help article "Advertiser guidance for allowing OpenAI web crawlers," as seen in search results, says:

  • You must allow OAI-AdsBot, and OpenAI recommends allowing OAI-SearchBot too.
  • Services such as Cloudflare or Akamai can block legitimate crawlers by mistake, often with a 403 Forbidden error. It asks you to allowlist OpenAI's crawler user agents in your WAF and CDN settings.
  • Human-verification and anti-automation checks in your own site can block the crawler too.
  • Login walls and rate limits can get in the way, so check those layers as well.

Again, the help center couldn't be read to confirm these details.

When you add an allow rule, match the OAI-AdsBot user agent and check requests against the IP ranges in openai.com/adsbot.json. Anyone can copy a user agent string, so the IP check stops fake bots from using your exception.

Step 4: Confirm OAI-AdsBot can load the page

Search your server or CDN logs for requests whose user agent contains OAI-AdsBot:

  • 200: it's getting through.
  • 403, 429 or a challenge page: a rule still blocks it. Your firewall's event log should name the rule.
  • No requests at all: OpenAI says it may visit the landing page when you submit an ad, and OAI-AdsBot only visits pages submitted as ads. Look again after Step 5.

Step 5: Get the ad reviewed again

Fixing the site doesn't change a review that already happened.

  1. Read the review result. Check the ad's review status and the reason given. In OpenAI's API these are review_status (in_review, rejected or approved) and review.
  2. Edit a rejected ad. OpenAI's docs say "Please edit your ad for it to be re-reviewed," and editing the creative "creates a new submitted version and starts another review."
  3. Don't just pause and restart. OpenAI says "Changing status does not approve a rejected creative or finish a pending review."

OpenAI's advertiser help page also says, in the version seen in search results, that ads may need to be re-uploaded or resubmitted once the landing page is reachable.

If your ad account itself is stuck in brand review, that's a different check: see how to sign up for ChatGPT Ads Manager and get approved. If the ad is approved but gets no impressions, see why a ChatGPT ad isn't getting impressions.

Common problems

ProblemWhy it happensFix
robots.txt has User-agent: * and Disallow: /Crawlers without their own group follow the * groupAdd a User-agent: OAI-AdsBot group with Allow: /
Your CDN blocks or challenges AI crawlersBot protection treats OAI-AdsBot like any other botAdd an allow rule for the OAI-AdsBot user agent, checked against openai.com/adsbot.json
The page loads for you, but the ad is rejectedYour browser passes bot checks; a crawler can't solve a challengeCheck your logs for OAI-AdsBot's responses (Step 4)
The ad points to a page behind a login, or to a link that isn't a web pageOAI-AdsBot can't sign in, and OpenAI requires an HTTP or HTTPS destination its ad crawlers can reachUse a public HTTP or HTTPS page
You fixed the site, but the ad is still rejectedThe old review standsEdit the ad to start a new review (Step 5)
You blocked GPTBot and worry about adsGPTBot is a separate user agentNothing to fix, unless a * rule or firewall blocks OAI-AdsBot too

Is this the same as Google's or Meta's ad crawlers?

Same idea, different crawlers and rules. For Google, see why Google Ads says "Destination not working", which covers firewalls that block Google's AdsBot. For Meta, see whether you can block Meta's crawler without breaking your ads. If you're still deciding whether to run ChatGPT ads at all, read whether ChatGPT ads are worth testing for B2B.

Sources (7)

Related resources