v0.1.7

WebGPU bridge update: llama.cpp b8157 baseline plus embedding APIs (embed/embedBatch).\n\nPublished assets: leehack/llama-web-bridge-assets@v0.1.7

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.

v0.1.2

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.

v0.1.1

Add bridge-owned CI and asset publishing workflows with validated emsdk setup.

v0.1.0

Initial public release of llama-web-bridge source (CMake, core bridge, JS runtime).