Download websites that use React Redux

About React Redux

Nobody sets out to adopt React Redux. The package is glue, picked up the moment a team decides a React application needs a central store, which is why one detection is really several facts arriving together. Maintained by Redux, React Redux connects a React component tree to a Redux store, which means the tag never travels alone. Where we detect it, the site is running React, running Redux, and has enough shared state that passing it down through components stopped being workable.

As of October 2026 we detect it on 2,491 sites. The JavaScript frameworks has been available since 2015 under the MIT licence, and its history tracks the history of the React ecosystem closely enough to work as a dating tool. Sites adopted it during the years when a central store was the standard answer for a large React application, while later projects increasingly chose lighter alternatives or React's own built in context, which makes a detection more likely to indicate an established application than a new build.

Version evidence supports that reading. Across installs where a version string survives the build we can read 12 distinct strings, and 7.2.6 recurs more often than any other. A long tail of versions on a library with a stable interface usually means the dependency gets upgraded when something else forces the issue rather than on its own merits. For anyone selling into these teams, a version string is a better indicator of when a front end was last rebuilt than of any opinion the team holds about state management.

Detection has a hard limit that should be stated plainly. The library is bundled at build time, so on a fully minified production build there may be nothing left in the page that identifies it. We record it where a source map, a separate vendor chunk, a development build or a recognisable module boundary exposes the name, which biases the recorded population towards sites that ship source maps or keep vendor bundles split out. The count is a floor rather than a total, and the version distribution leans towards installs that were legible rather than installs that were typical.

The practical case for a list rather than a lookup is the stack question. Anyone can open one site and read its bundle in a browser. What only the population answers is which companies in a market run a React application with centralised state, how those sites cluster by domain suffix, which of them pair the stack with a commerce platform, an analytics vendor or an error tracker you compete against, and which of them dropped the dependency between two of our snapshots. A migration signal of that last kind is invisible from a single lookup and obvious in a file, which is the whole reason to download the population instead of checking domains one at a time.