Government web census

Reports · fetched 2026-09-30 · source: CISA .gov registry

The government web census, September 2026: what 16,783 .gov domains publish for people, security researchers and AI agents

Every registered .gov domain in CISA's public list (16,783 domains: federal, state, county, city, tribal, school and special districts) was fetched once over HTTPS on 2026-09-30, with robots.txt, security.txt (both locations), llms.txt and the A2A agent card. 13,060 answered with a website; the rest are registered for email or internal use, are not yet live, or blocked our one polite request.

What we found

By level of government

LevelRegisteredReachablesecurity.txtvalidprivacy linkaccessibility linkHSTSblocks GPTBotllms.txt
Federal1,317992 (75.3%)74 (7.5%)4155.5%48.5%88.2%4318
State or territory1,483992 (66.9%)58 (5.8%)2038.0%38.2%44.4%4334
County3,0162,291 (76.0%)25 (1.1%)1330.2%40.7%30.4%105127
City9,4677,667 (81.0%)166 (2.2%)15332.2%39.1%28.9%314436
Tribal293191 (65.2%)2 (1.0%)118.3%4.2%37.2%414
School district7443 (58.1%)5 (11.6%)534.9%51.2%55.8%21
Special district1,102858 (77.9%)13 (1.5%)1041.4%36.5%47.4%1388

Federal domains without a security.txt (first 60 of 400)

acus.gov · arctic.gov · goldwaterinternship.gov · afrh.gov · arc.gov · ffiec.gov · access-board.gov · abmc.gov · fbiic.gov · whistleblower.gov · billetes.gov · americorps.gov · joinamericorps.gov · nationalservice.gov · mlkday.gov · goldwaterscholarship.gov · fedpartnership.gov · federalreserve.gov · cia.gov · smartcheck.gov · cftc.gov · asc.gov · mentor.gov · serve.gov · dnfsb.gov · aicenter.gov · presidentialserviceawards.gov · anchorit.gov · recalls.gov · aiexports.gov · saferproducts.gov · eda.gov · aviationweather.gov · pandemicoversight.gov · heat.gov · oasis.gov · dietaryguidelines.gov · nafri.gov · ars-grin.gov · usda.gov · farmerfairness.gov · symbols.gov · myplate.gov · safetyvideos.gov · bea.gov · mbda.gov · climate.gov · csb.gov · nwcg.gov · coldcaserecords.gov · uscurrency.gov · saferproduct.gov · firstnet.gov · lcacommons.gov · rural.gov · federalreserveconsumerhelp.gov · mgi.gov · xd.gov · poolsafely.gov · edison.gov

Government sites that tell ChatGPT and Claude not to open pages for users

aftonmn.gov · alton-tx.gov · appletonmn.gov · booneiowa.gov · atwatermn.gov · baraboowi.gov · belgradecitymn.gov · averytx.gov · belleplaineiowa.gov · beloitwi.gov · blackbrookwi.gov · bolingbrook.gov · brahammn.gov · bethlehemnj.gov · bradentonfl.gov · brockwaytwpmn.gov · braselton.gov · brunswickmd.gov · brysoncitync.gov · camarillo.gov · byesvilleoh.gov · canyonlakeca.gov · cedarpointnc.gov · charlestownri.gov · chattahoocheefl.gov · centralcityia.gov · cedartowngeorgia.gov · chathamnewyork.gov · chiltonwi.gov · circlepinesmn.gov · cityofbensonmn.gov · cityofbyronmn.gov · cityofclaremontmn.gov · cityofclearlakemn.gov · cityofcorningny.gov · cityofclevelandmn.gov · cityofguttenbergia.gov · cityofkenyon.gov · cityoflagrangemo.gov · cityofmarionwi.gov

Government sites publishing an llms.txt

adriantownshipwi.gov · adriantwpmi.gov · alberttwpmi.gov · almacenterwi.gov · abitaspringsla.gov · amalgautah.gov · altontownshipil.gov · alvo.gov · argonnewi.gov · barringtontwpil.gov · annettatx.gov · aspinwallpa.gov · ashtonid.gov · arroyopr.gov · bearriverwy.gov · beavertwp-oh.gov · belgiumwi.gov · auburnmi.gov · beechgrove.gov · assyriatwpmi.gov · birchlaketownshipmn.gov · ausabletwp-mi.gov · bishophillil.gov · bluefieldwv.gov · bethelparkpa.gov · arcadami.gov · bloomingdaletownshipmi.gov · bloomingdalevillagemi.gov · bogotanj.gov · bogotapolicenj.gov · blanding-ut.gov · bloomingdale-ga.gov · brackettvilletx.gov · bradfordme.gov · bridgeportne.gov · brinkhaven.gov · bristoltownshipohio.gov · brooklyntownwashburnwi.gov · bucklinmo.gov · burnham-il.gov

Look up any .gov domain

Method and limits

Population: CISA dotgov-data current-full.csv (all registered .gov domains with type, organization, city and state). Fetch: one HTTPS request to the apex (then www.), following redirects, 8-second timeout, user agent agentexchange-census/1.0; robots.txt, /.well-known/security.txt, /security.txt, /llms.txt and /.well-known/agent-card.json on the final host. "Reachable" means any HTTP response over HTTPS. Homepage links are detected from anchor targets and visible text, so a policy that exists but is not linked from the homepage is not counted. A robots.txt "blocked" verdict means the agent's group (or *) disallows the whole site. One snapshot; municipal sites that were slow or that block automated requests are counted as unreachable. Data: JSON (CC BY 4.0).

Run a .gov site? Generate a security.txt robots.txt rules for AI crawlers · llms.txt generator · accessibility feedback desk (free)