Robots · User-Agents · IP address

Our robots

If you found globetrotters.ai/bot in your logs, one of our robots visited your site. They act for travel brands that use Globetrotters to manage their Official AI Presence, and for anyone who runs the free AI Readiness audit. They read public pages only, and make few requests.

Our egress IP address

34.53.138.166

One static IPv4 address. We do not send requests over IPv6.

Our User-Agents

Every request from our robots carries one of these tokens, with a link back to this page.

AI Readiness audit

GlobetrottersReadinessBot/1.0 (+https://globetrotters.ai/bot)

Reads a website the way an AI assistant would and scores how well it can understand it. The audit runs when the site's owner requests it in Studio, or when anyone audits the site on readiness.globetrotters.ai.

The token is appended to a desktop Chrome User-Agent, so pages render as they would for a browser:

Mozilla/5.0 (Macintosh; Intel Mac OS X 10_15_7) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/130.0.0.0 Safari/537.36 GlobetrottersReadinessBot/1.0 (+https://globetrotters.ai/bot)

Install check

GlobetrottersApexProbe/1.0 (+https://globetrotters.ai/bot)

Checks that an Official AI Presence installed on your own domain (WordPress plugin or Symfony bundle) is live and up to date, when you click Check install in Studio, and verifies your IndexNow key. One or two requests per check.

One of those requests tests whether your site lets GPTBot through. It uses the GPTBot User-Agent with our token appended:

Mozilla/5.0 (compatible; GPTBot/1.0; +https://openai.com/gptbot) GlobetrottersApexProbe/1.0

Favicon

GlobetrottersFaviconBot/1.0 (+https://globetrotters.ai/bot)

Fetches a brand's favicon or logo once, to show it on its presence pages.

The audit also checks how your site answers the AI crawlers themselves (GPTBot, ClaudeBot, PerplexityBot…). Those requests go through a third-party rendering service and use the crawlers' own User-Agents, so they do not come from our IP address.

How much we request

~25

requests per audit, over a few seconds

5

requests in parallel, at most

250 ms

at least, between requests for files that may not exist

1×

each URL is requested once per audit

Allow our robots

Some tools ban an IP address after a few 404s in a row, which is exactly what an audit looking for /.well-known/ files produces. If you use one of them, allow our address. Allowing by IP is more reliable than by User-Agent: most ban tools only see the IP.

fail2ban

Add our address to ignoreip in jail.local, under [DEFAULT] to cover every jail. A jail's own ignoreip replaces the default rather than adding to it, so repeat the address there if you set one.

# /etc/fail2ban/jail.local
[DEFAULT]
ignoreip = 127.0.0.1/8 ::1 34.53.138.166

Reload, and lift a ban already in place:

sudo fail2ban-client reload
sudo fail2ban-client unban 34.53.138.166

Prefer ignoreip to an ignoreregex on our User-Agent: most jails (apache-botsearch among them) read log lines that do not contain the User-Agent, so the regex never matches.

CrowdSec

On CrowdSec 1.6.8 or later, add our address to an allowlist. It applies at once:

sudo cscli allowlists create globetrotters -d "Globetrotters robots"
sudo cscli allowlists add globetrotters 34.53.138.166 -d "https://globetrotters.ai/bot"

On earlier versions, use a parser whitelist:

# /etc/crowdsec/parsers/s02-enrich/globetrotters-whitelist.yaml
name: globetrotters/whitelist
description: "Globetrotters robots - https://globetrotters.ai/bot"
whitelist:
  reason: "Globetrotters robots"
  ip:
    - "34.53.138.166"

Then reload, and remove a decision already in place:

sudo systemctl reload crowdsec
sudo cscli decisions delete --ip 34.53.138.166

Apache 2.4

With ModSecurity, turn the rule engine off for our address. Place the rule before your rule set is loaded (with the OWASP CRS, in REQUEST-900-EXCLUSION-RULES-BEFORE-CRS.conf), and pick an id that is free on your server:

SecRule REMOTE_ADDR "@ipMatch 34.53.138.166" \
    "id:1000001,phase:1,pass,nolog,ctl:ruleEngine=Off"

With mod_evasive, add it to the module's configuration:

DOSWhitelist 34.53.138.166

For a location restricted by IP, add it next to your existing rules:

<Location "/private">
    <RequireAny>
        Require ip 34.53.138.166
        # ...your existing Require lines
    </RequireAny>
</Location>

nginx

In a location restricted by IP, allow our address before the deny:

location /private/ {
    allow 34.53.138.166;
    # ...your existing allow lines
    deny all;
}

Cloudflare

Create a WAF custom rule (Security → WAF → Custom rules) with this expression and the action Skip, skipping the remaining custom rules, rate limiting rules, managed rules and Super Bot Fight Mode:

(ip.src eq 34.53.138.166)

Bot Fight Mode on the Free plan cannot be skipped by a custom rule. Add an IP Access rule for our address with the action Allow instead (Security → WAF → Tools).

Contact

A question about our robots, or a request you do not recognise? Write to contact@globetrotters.ai with your domain and, if you can, a log line.