For site owners and developers
About the Echobat scanner
Our scanner visited your website and you went looking for us. This page says who we are, what we read, and how to let us in or send us away. The short version is on the EchobatBot page.
Who we are
Echobat tells a local business how it looks to AI assistants. Someone asks ChatGPT or Google’s AI for a plumber nearby and an answer comes back naming three businesses. We read a website the way the crawlers behind those assistants read it, and report what we find to the business that owns it.
We are not a search engine, we do not sell data, and we do not build a database of other people’s content. A scan reads a handful of pages, grades eleven things about them (ten measured automatically, one checked by hand), and the result goes to one customer: the person who asked for it.
Why we read your site
One of three reasons, and you can tell which from your logs:
- Someone asked us to. Almost always the business itself, or the person they pay to look after the site, typing their own address into our scanner.
- Someone is comparing. An agency or a competitor can run a scan on any public address, the same way they could open the site in a browser.
- A monthly re-scan. A paying customer gets their site re-read once a month so they find out when something slips.
We read up to 15 pages per scan, one at a time, and reading them takes about a minute. We do not run JavaScript, we do not submit forms, we do not log in, and we do not touch anything behind a password. We read the HTML your server sends, plus robots.txt and your sitemap.
We obey robots.txt
We read robots.txt before anything else and we honor it. If it disallows us, we stop there and fetch no pages at all. The customer who asked for the scan is told the site declined, and shown your file as the reason. If it disallows part of the site, we skip that part and read the rest.
We follow the rules written for EchobatBot by name where you have written any (rules for our old name, EchobatScanner, still count), and the User-agent: * rules otherwise. Wildcards and $ anchors work as you would expect, the longest matching rule wins, and an Allow beats a Disallow of the same length, which is the reading that favors your intent.
Letting us in
Our scanner identifies itself honestly on every request. This is the exact string:
EchobatBot/1.0 (+https://echobat-ai.com/bot)
Requests are also signed with Web Bot Auth, so they can be checked against our published key. We never pretend to be a browser. That is deliberate: the whole point of Echobat is to report what a site does to a non-browser visitor, and a scanner that disguised itself would come back with a clean result while the real AI crawlers were still being turned away. It also means a firewall rule that blocks “anything that is not a browser” blocks us, and probably blocks them too.
Cloudflare
- Security → Events. Find the blocked request first and note which feature stopped it. Only change the one that is actually firing.
- Security → WAF → Custom rules → Create rule.
- Expression:
(http.user_agent contains "EchobatBot") - Action: Skip, ticking at least All remaining custom rules. Add All managed rules and All Super Bot Fight Mode rules if those are what stopped us.
- Put the rule above any stricter rule that would match the same traffic: custom rules run in order, and a Block ends evaluation.
One thing worth knowing: if your site is on Cloudflare’s Free plan you probably have classic Bot Fight Mode, and Cloudflare’s own documentation says that one cannot be skipped by a custom rule. The rule above will not help against it. Your options there are to turn that feature off for the zone, or to leave us blocked, which is a legitimate choice, and the next section says how to make it permanent.
Wordfence
Wordfence has no simple “allow this user-agent” setting, so this is diagnosis first:
- Wordfence → Tools → Live Traffic, filtered to blocked requests. Find our user-agent and read the rule that fired. That tells you which of the next steps you need.
- Blocked as a known bad user-agent? That is Wordfence’s built-in protection. There is no per-agent exception for it, and we would not advise relaxing it on our account.
- Blocked by rate limiting? Wordfence → Firewall → Rate Limiting. The settings that catch a scanner are “If a crawler’s page views exceed” and “If anyone’s requests exceed”. We read up to 15 pages in about a minute, so a limit below roughly 30 page views a minute will stop us. Setting the action to Throttle rather than Block is gentler on every legitimate crawler, not just ours.
- Check for an old custom rule. Wordfence → Firewall → Blocking may contain patterns a previous developer added that match crawler user-agents or whole datacenter IP ranges. Those often predate AI crawlers entirely and are worth a look regardless of us.
Asking us to stop
The quickest way is robots.txt. We obey it, and it takes effect on our next visit without anyone having to answer an email:
User-agent: EchobatBot Disallow: /
Or email support@echobat-ai.com with the domain and we will stop scanning it. No form, no account, no conditions, and a person reads it. That applies to the whole domain for every customer, including the business itself, so if you are the owner and you would rather not be scanned at all, it is the same email.
If a scan is causing a problem
Email support@echobat-ai.com and say what you saw. We would rather hear about it. One page per request, one request at a time, and no scan of the same site twice within 24 hours. If your logs show something heavier than that, something is wrong on our end and we want to know.
Contact
support@echobat-ai.com
Echobat, P.O. Box 24381, Winston-Salem, NC 27114