HTTP/3 is the third major version of the HTTP protocol, standardized as RFC 9114 in June 2022. It carries the same HTTP semantics as earlier versions, but sends them over QUIC, a transport protocol built on UDP, instead of TCP.
How does HTTP/3 differ from HTTP/2?
With HTTP/2, many streams share one TCP connection, so a lost packet delays all active streams. QUIC provides multiplexing itself. If a packet is lost, only the streams whose data was in that packet wait. HTTP/3 also replaces the HPACK header compression of HTTP/2 with QPACK, which works with QUIC’s out-of-order delivery. QUIC includes TLS encryption, so HTTP/3 always runs encrypted.
| Version | Transport | Concurrent requests | Header fields | Use by Google's crawlers |
|---|---|---|---|---|
| HTTP/1.1 (RFC 9112) | TCP | no multiplexing layer; pipelining possible, responses come back in request order; often several TCP connections for parallel requests | whitespace-delimited text fields; often repetitive and verbose according to RFC 9113 | supported; Google's default version |
| HTTP/2 (RFC 9113) | TCP | yes, many requests interleaved on one connection; TCP-level packet loss still stalls all streams | binary frames, HPACK compression | supported since November 2020; no ranking boost; opt out with status code 421 |
| HTTP/3 (RFC 9114) (this term) | QUIC over UDP | yes, over independent QUIC streams; one blocked stream does not hold up the others | QPACK compression | not stated by Google; the crawler documentation lists only HTTP/1.1 and HTTP/2 |
All three versions carry the same HTTP semantics (methods, status codes, headers); they differ in how messages are transferred. Google's crawlers pick the version with the best crawling performance and may switch between crawling sessions. Google says HTTP/2 may save computing resources on both sides. Server logs show which version a request used. Sources: RFC 9112: HTTP/1.1; RFC 9113: HTTP/2; RFC 9114: HTTP/3; RFC 9000: QUIC; Google Crawling Infrastructure: Overview of Google crawlers and fetchers (last updated 2026-06-12); Google Search Central Blog: Googlebot will soon speak HTTP/2 (2020-09-17). Last checked: .
How do clients find HTTP/3?
A client usually starts with HTTP/1.1 or HTTP/2 over TCP. A server announces HTTP/3 with the Alt-Svc response header, for example Alt-Svc: h3=":443", and the client can use QUIC for later requests. A DNS record of type HTTPS (RFC 9460) lets clients learn about HTTP/3 before the first connection. If a UDP connection fails, clients fall back to a TCP-based version.
HTTP/3 and crawlers
Google’s overview of its crawlers and fetchers states support for HTTP/1.1 and HTTP/2. HTTP/3 is not mentioned there. Enabling HTTP/3 on a server or CDN therefore changes how browsers connect, and no effect on Google’s crawling is documented. Server logs show which protocol each request used.
Example
HTTP/2 200
alt-svc: h3=":443"; ma=86400
The ma value is how long, in seconds, the client may remember the alternative service.
Sources
- RFC 9114: HTTP/3rfc-editor.org
- HTTP/3 – Wikipediaen.wikipedia.org
- Overview of Google crawlers and fetchers – Google Search Centraldevelopers.google.com
Last updated: