AcroBot

• Site Audit: crawls up to 2,000 pages of a site (10,000 at most) and checks titles, links, redirects, speed signals, structured data, robots.txt, sitemaps, llms.txt and whether AI assistants can read the site.
• Local SEO check: compares the website's name, address, phone and hours with the business's Google profile.
• Broken-link check: sends a light HEAD/GET request to links that point to other sites, so it can report broken outbound links.
• Ownership check: fetches one file (/acrobot-verify-<token>.html) or reads a DNS TXT record (acrobot-verification=<token>) when a site owner verifies their site with Acro.

AcroBot does not build a public index, sell data, train AI models, fill in forms or try to log in. By default it does not follow nofollow links.

robots.txt token: AcroBot, matched case-insensitively.

A site owner who has verified ownership with Acro can ask for an audit with a standard Chrome User-Agent. That is the only case where we crawl without the AcroBot name, and it is always from the IPs below.

AcroBot only crawls from these addresses. Requests that claim to be AcroBot from any other IP are not from us.

A machine-readable copy is at https://getacro.ai/bot/ips.json. Firewalls and scripts can read it the way they read Google's googlebot.json.

{
  "updated": "2026-10-24",
  "prefixes": [
    {
      "ipv4Prefix": "3.6.24.92/32"
    }
  ]
}

Unlike crawlers that rotate through large IP ranges, AcroBot only crawls sites a customer asked us to audit, so it uses a short, fixed list you can trust and allow-list.

• Starts at the homepage and the sitemaps, and follows links on the same site.
• Reads robots.txt at the start of every audit and obeys the rules for AcroBot, or * when there is no AcroBot group.
• By default it crawls at most 2 requests per second per site (120 URLs per minute). It halves its speed on 429/503 or slow responses, honours Retry-After, and never drops below one request every 2 seconds. An Acro customer can raise the limit to 5 requests per second; anything faster needs the site owner to verify ownership.
• One crawl at a time per site, across all of Acro's servers.
• Audits run when a customer starts one, and on a schedule (weekly by default).
• Never crawls private or internal network addresses.

Add these rules to robots.txt in the site root, and on each subdomain.

# Block AcroBot completely
User-agent: AcroBot
Disallow: /

# Or block only some folders
User-agent: AcroBot
Disallow: /private/
Disallow: /checkout/

Changes apply from the next audit, because robots.txt is read at the start of each one. If robots.txt is missing (HTTP 404), AcroBot crawls normally. If it returns a server error (5xx) or can't be reached, AcroBot treats the whole site as off-limits and stops.

If your firewall or host blocks AcroBot, the audit stops or reports false errors. Allow-list these IPs:

1. The IP is in the list above (or in ips.json).
2. Optional: host <IP> returns crawl-N.bot.getacro.ai, and host crawl-N.bot.getacro.ai returns the same IP.