The encyclopedia · Software & IT · Technical decision · 2015–2018
Brotli made web pages smaller and shipped the decoder in Chrome
Google's LZ77-plus-context-model codec beat gzip by ~20% and arrived already inside the browser
the move
HTTP compression had effectively frozen at gzip for two decades: every browser and server spoke it, but its 1992-era design left obvious headroom. Google's Jyrki Alakuijala and Zoltan Szabadka designed Brotli to close that gap.
RFC 7932 defines Brotli as a lossless format combining LZ77 back-references and Huffman coding with context modeling and a static dictionary, giving compression efficiency comparable to the best general-purpose methods of its time — roughly 20% smaller than gzip on typical web text.
As with WebP, the adoption trick was distribution: Brotli's decoder shipped inside Chrome and later Firefox and Safari, so web servers could start serving Brotli the moment clients announced it. MDN documents it as a standard end-to-end HTTP compression option, and CDNs, WordPress, and nginx all adopted it.
why it works
- A browser-native decoder meant zero friction for visitors and instant demand-side support
- The RFC specified the format precisely, so independent encoders and decoders interoperate
- Static dictionaries tuned for web text gave real-world gains beyond generic LZ77
- Falling CPU costs made the heavier encoding affordable for servers, while clients decode cheaply
what transfers
A replacement wins when the client side is handled: publish the spec, ship the decoder in the dominant browser, and servers adopt overnight.
what came after
Brotli became the default text-compression choice across major CDNs, web frameworks, and static hosts, and its successor work continued inside the same Google compression group. Gzip remained the baseline, but Brotli defined the modern ceiling for HTTP transfer size.
references
spotted an error? The archive wants to know.