Download websites that use Phoenix LiveView

About Phoenix LiveView

A rendering architecture is one of the more revealing entries in a technology catalogue, because it encodes a bet a team made about where application state should live. Phoenix LiveView keeps that state on the server, renders HTML there, and pushes small diffs to the browser across a persistent connection, so the page behaves like a single page application without anyone writing one. Adopting it at any point since 2019 means a team decided against maintaining a separate front end codebase.

Consequences of that bet are what make the population interesting. A deployment holds an open connection for every active user, which places real demands on the infrastructure in front of it. Socket aware load balancing, routing that keeps a session on the same node, a CDN configuration that will not terminate long lived connections, and capacity planning measured in concurrent sessions rather than in requests. Released under the MIT licence and written in Elixir, the library also implies a small front end team or none at all, and a hiring profile weighted toward backend engineers.

As of October 2026 we record 1,377 sites, and one caution matters more here than the usual ones. A site running this library is by construction also running the framework it belongs to and the runtime beneath that, so three entries in our catalogue describe heavily overlapping populations whose counts must never be added together. Detection also depends on the client script being served in a recognizable form, and a build pipeline that bundles it with everything else erases the evidence, so the visible set skews toward conventional asset pipelines.

Practical uses follow directly from what the architecture implies. Infrastructure vendors selling connection aware hosting, load balancing, or real time monitoring get a list of companies that already have the problem they solve. Competing front end frameworks and component vendors get a list of teams that chose against them, which is more informative than a list of teams that chose them. Researchers tracking whether server driven interfaces are a genuine shift or a passing niche can measure it across repeated snapshots instead of arguing from conference talks. And a company evaluating the same architecture can see who runs it in production, in which industries and at what apparent scale, before committing to it. A lookup on one domain answers none of those, because every one of them is a question about a set.