Google Lists 2,181 Crawler IP Prefixes; a Legacy File Still Serves a March Copy Missing 9 Ranges

Ras Al Khaimah, United Arab Emirates, 17 October 2026.

Google's five published IP range files for its crawlers and fetchers listed 2,181 network prefixes on 17 October 2026, covering 104,608 IPv4 addresses, according to a SaaS Link Building analysis of the files Google links from its Verify requests from Google crawlers and fetchers page. The analysis also found that one file at Google's old /search/apis/ipranges/ location still serves a copy dated 3 March 2026 that lacks nine IPv4 ranges found in the current version, more than six months after Google announced the move.

Sites that allowlist or verify Googlebot by IP, rather than by reverse DNS, depend on these files being current.

Key findings

• The five files list 2,181 prefixes in total: 1,104 IPv4 and 1,077 IPv6.
• User-triggered fetchers running on Google Cloud make up the largest file, with 1,068 prefixes, nearly half of the total.
• The common crawlers file, which covers Googlebot, lists 317 prefixes. 122 of its 170 IPv4 prefixes sit inside 66.249.0.0/16, the block most SEOs associate with Googlebot.
• All five current files were regenerated on 16 October 2026, the day before our pull.
• Four of the five legacy addresses under /search/apis/ipranges/ now return a 301 redirect to the new /crawling/ipranges/ location. The fifth, user-triggered-agents.json, returns HTTP 200 with a March copy.

File	Total prefixes	IPv4	IPv6	IPv4 addresses	Generated
common-crawlers.json (Googlebot and other common crawlers)	317	170	147	5,088	16 Oct 2026
special-crawlers.json (for example AdsBot)	272	136	136	4,352	16 Oct 2026
user-triggered-fetchers.json (Google Cloud, gae.googleusercontent.com)	1,068	534	534	17,088	16 Oct 2026
user-triggered-fetchers-google.json (google.com hostnames)	504	252	252	8,064	16 Oct 2026
user-triggered-agents.json	20	12	8	70,016	16 Oct 2026
Total	2,181	1,104	1,077	104,608

The stale legacy file

On 31 March 2026 Google announced on the Search Central blog that the JSON files were moving from /search/apis/ipranges/ to /crawling/ipranges/, because the ranges apply to more than Search crawlers. The post, signed by Gary, said the old files would stay available for a while and that Google would phase out the old locations and redirect them to the new ones within six months.

Six and a half months later, the redirect is in place for four files. The old googlebot.json now redirects to common-crawlers.json, and the special crawlers and both user-triggered fetcher files redirect to their new copies. The old user-triggered-agents.json does not. It returned HTTP 200 with a Last-Modified header of 3 March 2026 and a creationTime of the same date.

user-triggered-agents.json	Legacy /search/ path	Current /crawling/ path
Generated	3 Mar 2026	16 Oct 2026
Prefixes	4	20
IPv4 prefixes	3	12
IPv6 prefixes	1 (/48)	8 (/124)
IPv4 ranges missing from legacy copy	9 ranges, 2,176 addresses (136.121.40.0/21 and eight /28 blocks in 74.125.232.0/24)

The IPv6 picture runs the other way. The legacy file lists a single /48, which contains all eight /124 blocks in the current file, so a system using the old file would accept more IPv6 addresses than Google now lists. A system using the old file for IPv4 would reject requests from the nine newer ranges.

Google's verification page lists user-triggered-agents.json alongside the two user-triggered fetcher files and says user-triggered fetchers ignore robots.txt because a user asked for the fetch. Google has not said publicly why this one legacy address was left without a redirect.

Other patterns in the data

Every prefix in the four main files sits inside Google's general list of IP addresses at gstatic.com/ipranges/goog.json. In the user-triggered agents file, the four prefixes in 136.121.0.0/16 and 136.122.0.0/16 do not fall inside any range in that general list, so a check against goog.json alone would miss them.

Prefix sizes are uniform in most files. Every IPv4 prefix in the special crawlers and both user-triggered fetcher files is a /27, and every IPv6 prefix is a /64. The common crawlers file mixes 148 /27 blocks with 22 /28 blocks. The user-triggered agents file is the outlier: it includes a full /16 (65,536 addresses), which is why a file with 20 prefixes accounts for 67% of all listed IPv4 addresses.

The common crawlers file is not limited to 66.249.0.0/16. Its IPv4 prefixes span 23 different /16 blocks, including 25 prefixes in 192.178.0.0/16 and others in 34.x.x.x ranges such as 34.118.0.0/16 and 34.89.0.0/16. Google's verification page shows an example geo-crawl hostname resolving to a 35.x address, which fits that spread.

What this means for SEO teams

• Update any hardcoded URLs. If a firewall, CDN rule or log parser still fetches from /search/apis/ipranges/, switch it to /crawling/ipranges/. For user-triggered-agents.json the old address is serving data that is seven months out of date.
• Refresh on a schedule. All five current files were regenerated the day before our pull. A weekly or daily refresh is safer than a one-off import.
• Use the right file for the job. Googlebot sits in common-crawlers.json. Allowlisting all five files also lets through user-triggered fetchers that do not follow robots.txt.
• Keep DNS verification as a fallback. Google's documentation still describes a reverse and forward DNS lookup, with hostnames ending in googlebot.com, google.com or googleusercontent.com, as enough for most cases.
• Do not rely on 66.249 alone. More than a quarter of Googlebot's IPv4 prefixes fall outside that block.

Methodology

On 17 October 2026 at about 03:00 UTC we downloaded the five JSON files linked from Google's verification page (common-crawlers, special-crawlers, user-triggered-fetchers, user-triggered-fetchers-google and user-triggered-agents) from developers.google.com/static/crawling/ipranges/, plus goog.json from gstatic.com. We requested the five matching file names under the legacy path developers.google.com/static/search/apis/ipranges/ and recorded each HTTP status and redirect target. We parsed each file with Python's ipaddress module, counted ipv4Prefix and ipv6Prefix entries, summed IPv4 addresses per prefix, grouped IPv4 prefixes by their first two octets, and tested each prefix for containment in goog.json and in the legacy file. Address totals count every address in each CIDR block and do not remove overlaps between files; we found no identical prefix listed in more than one file. The files change often, so later pulls will differ.

Conclusion

Google currently publishes 2,181 crawler and fetcher prefixes across five files, and all five were refreshed on 16 October. The move to /crawling/ipranges/ is mostly complete, with redirects on four of the five old addresses.

The exception matters for anyone who verifies Google traffic by IP: the legacy user-triggered-agents.json still serves a March copy that is missing nine IPv4 ranges. Teams should point their tooling at the new location and refresh it regularly.

Questions and answers

Where are Google's crawler IP range files now?

Google moved them to developers.google.com/static/crawling/ipranges/. A 31 March 2026 Search Central blog post said the old /search/apis/ipranges/ location would be redirected within six months.

Which file contains Googlebot's IP addresses?

Googlebot is in common-crawlers.json. On 17 October 2026 it listed 317 prefixes: 170 IPv4 and 147 IPv6. The old googlebot.json address now redirects to it.

Which legacy file is out of date?

The legacy user-triggered-agents.json under /search/apis/ipranges/ returned HTTP 200 on 17 October 2026 with a 3 March 2026 copy of 4 prefixes. The current file at the new path lists 20 prefixes, including nine IPv4 ranges not in the old copy.

Do Google's user-triggered fetchers follow robots.txt?

No. Google's verification documentation says user-triggered fetchers ignore robots.txt rules because the fetch was requested by a user.

###