RepoWatch / GitHub signal

Llama.cpp b9672 Release

Core local inference engine updated; bump your stacks.

Foundational for Hermes, OpenClaw local agents and any llama.cpp-based tooling.

What changed

  • Release: b9672 (from b9663)
  • Commit: UI fix for SSE transport detection and routing through CORS proxy.
  • Downstream: ollama/ollama updated to b9672 in v0.30.9 release.

Why it matters

llama.cpp remains the go-to engine for efficient local inference. The SSE/CORS fix improves reliability for web UIs and agent frontends. Ollama’s quick follow-up means most users get it via the popular wrapper.

Relevant to local-inference-optimisation category and our agent tooling.

My read

Solid incremental release. The UI transport fix addresses real pain points in browser-based or proxied setups. Not revolutionary but worth pulling in for stability.

Bottom line

Update now if running llama.cpp or Ollama in production stacks. Worth a spike for any local agent work.