Boot a kernel.

A real Linux kernel (riscv64, container2wasm) boots inside this tab. Then spawn pods: shells with their own PID, mount, UTS, IPC and network namespaces and a private copy-on-write root, all on that one kernel. Everything you see in the topology is read back from /proc, not from what was asked for.

The guest gets a network adapter whose gateway is this page: pods get their own adapters on a bridge, an ACL decides every flow, and a driver carries what is allowed out of the tab (fetch for HTTP, a WebSocket relay for everything). The ⌁ network tab has the controls.

Nothing leaves the browser. A real image is 60 to 140 MB, fetched once and cached; the simulation is a JavaScript model speaking the same protocol.

Then: name a pod, tick what it should get, Spawn pod. Type echo $$ in it. Press ? for the tour, or read how it works.