CrawlCheck

Guides · 2026-10-18 · By · 0 views

HTTP/3: how to check if your site offers it, and who uses it

A site tells browsers it speaks HTTP/3 in two places: a response header and a DNS record. Here is how to read both, what three real sites showed, and why search and AI crawlers do not use HTTP/3 yet.

A site offers HTTP/3 if it advertises h3 in the Alt-Svc response header (RFC 7838) or in a DNS HTTPS record with alpn=h3 (RFC 9460). Check the header with curl -sI and the record with a DNS lookup for type HTTPS. On 18 October 2026 google.com advertised h3 in both, crawlcheck.io only in DNS, and nps.gov's CDN offered HTTP/2 only. Google's crawlers support HTTP/1.1 and HTTP/2, not HTTP/3, so HTTP/3 helps visitors on slow networks rather than crawling.

Check your own site free

About a minute. No account, no card, nothing installed.

HTTP/3 (RFC 9114) carries web requests over QUIC, which runs on UDP datagrams instead of a TCP connection. Its practical advantage is on lossy networks, such as a phone moving between cell towers: a lost packet holds up only the request it belongs to, not every request on the connection. Browsers use it when a site says it is available. This guide shows where a site says so and how to check.

The two places a site advertises HTTP/3 #

WhereWhat to look forStandard
Response headeralt-svc: h3=":443"; ma=86400RFC 7838
DNS HTTPS recordalpn=h3,h2 in the recordRFC 9460

The header is learned after a first visit over HTTP/2 or HTTP/1.1; the browser remembers it for the ma (max age) in seconds. The DNS record is read before the first connection, so a browser can start on HTTP/3 straight away. The same DNS record can also carry an ech= key for Encrypted Client Hello, which hides which site a visitor is connecting to.

How to check your site #

  1. The header: curl -sI https://example.com/ | grep -i alt-svc. No line means the header does not advertise HTTP/3.
  2. The DNS record: dig example.com HTTPS +short, or over HTTPS: curl -s -H 'accept: application/dns-json' 'https://cloudflare-dns.com/dns-query?name=example.com&type=HTTPS'. Look for alpn= with h3 in it.
  3. Check the host you actually serve. If the bare domain redirects to www, check www, and follow any CNAME to your CDN: the record that counts is the one at the end of the chain.
  4. Or run the free scan: every report now has an HTTP/3 advertised row in the hosting section.

Three real sites #

SiteAlt-Svc headerDNS HTTPS recordResult
google.comh3alpn includes h3Advertised in both
crawlcheck.iononealpn=h3,h2, with echAdvertised in DNS only
nps.gov (www, on Amazon CloudFront)nonealpn=h2Not advertised

All three checked on 18 October 2026. Our own site gets its DNS record from Cloudflare, which publishes HTTP/3 and Encrypted Client Hello there, but its responses do not send the Alt-Svc header. A browser that reads the DNS record still finds HTTP/3; one that relies on the header does not. The nps.gov result is not a fault: its CDN record offers HTTP/2, which is a reasonable choice.

Do crawlers use HTTP/3? #

Not Google's. Google states that its crawlers and fetchers support HTTP/1.1 and HTTP/2, use HTTP/1.1 by default, and pick whichever performs better per session; HTTP/3 is not mentioned (Google's crawler overview). We have not seen an AI crawler document HTTP/3 support either. So HTTP/3 does not change how a site is crawled, indexed or cited today. It changes how fast it feels to people.

What the check cannot tell you #

Advertising HTTP/3 is not the same as serving it well. A server-side check reads the header and the DNS record; it cannot complete a QUIC handshake, so it cannot prove that UDP port 443 is open end to end. Some corporate networks block UDP, and browsers then fall back to HTTP/2 silently. That is why CrawlCheck reports the HTTP/3 row without scoring it.

Should you turn it on? #

If your CDN offers it, usually yes: it costs nothing for crawlers, which keep using HTTP/2 or HTTP/1.1, and helps visitors on mobile networks. Make sure the advertisement and the reality agree. A site that advertises HTTP/3 and then drops QUIC traffic makes browsers wait before falling back. For what crawlers do need, start with the guide to checking AI crawler access.

Every figure above came out of CrawlCheck.

Point it at your own domain and see the same measurements, free.

Scan a domain — free

The main product

Found this on your own site? We fix it for $500.

Scan free to see where you stand. The fix is one site, every finding implemented and re-measured, with a sealed before and after.

Fix it all — $500

What it covers · scan first, it is free

Related findings

2026-10-02IndexNow: what it does, who reads it, how to check itIndexNow tells engines a URL changed. One submission reaches Bing, Yandex, Seznam, Naver, Yep and others; Google does not take par2026-10-04402 Payment Required: what top sites send AI crawlersWe read the top 1,000 sites as four AI crawlers. Three answered with 402 Payment Required and none quoted a price: a refusal, a to2026-08-13Which AI crawlers actually visit a small business siteNot a probe: real, unasked-for visits to sites we operate, each checked against the operator's published IP ranges. Which AI crawl
How anything measured in this article was measured15client identitiesone second, one address5machine filesapex and www114named agentsresolved from robots.txt24sections scoredreach, read, quoteHow anything measured here was measured15 client identities5 machine files114 named agents26 sections scoredone second, one addressapex and wwwresolved from robots.txtreach, read, quote
No account, nothing installed, and the same sequence on every domain — which is what makes one scan comparable to another. Run it on your own site.

Comments

Comments are read before they appear. Nothing is published automatically, and no account is needed.

Writing about this? Facts, live figures and marks — every number on that page is dated and traceable to a scan.

All guides · The dataset · How the dataset works