Native and browser runtime boundaries

Native and browser execution share model concepts but differ in memory, threading, storage, device access, policy, and failure behavior.

Browser and native architectures should not be treated as interchangeable deployment targets. The loader, memory manager, backend, and cache may implement the same ports while obeying different host constraints.

Native hosts

Native processes may use filesystem mappings, operating-system threads, accelerator drivers, and host-level observability. These capabilities still require version and platform verification.

Browser hosts

Browser research must account for sandboxed storage, worker availability, memory ceilings, cross-origin isolation, browser GPU support, and user-agent-specific behavior.

Website boundary

The MiRust WordPress website documents these topics. It ships no Rust, WebAssembly, WebGPU inference, model weights, or browser execution path.