v0.1.5

What changed

• Add resilient large-model web loading in the bridge runtime, including streamed staging with resume/retry support and redirect-aware remote URL handling.
• Enable pthread-capable bridge builds and ship a dedicated worker entry so fetch-backed loading works in cross-origin-isolated browser contexts.
• Add wasm64 build support improvements (memory64 options, 12 GiB default cap, wasmfs fetch patch) and update build/runtime documentation.

Note

• This release is intended to align downstream asset versioning with the existing llama-web-bridge-assets sequence.