llama.cpp v10010 adds CORS options for local AI servers
New release brings cross-origin support to local LLM inference…
ggml-org’s llama.cpp project, a popular C++ implementation for running large language models locally, issued version b10010 on July 14. The release introduces new --cors-* CLI options (e.g., --cors-allow-origin, --cors-allow-methods) to control Cross-Origin Resource Sharing behavior. This allows developers to securely serve llama.cpp’s HTTP interface from web apps hosted on different domains, a common requirement for modern AI tooling.
The release ships precompiled binaries for a wide range of platforms: macOS (Apple Silicon & Intel), Linux (x64/arm64 with CPU, Vulkan, ROCm, OpenVINO, SYCL), Windows (x64/arm64 CPU, CUDA 12/13, Vulkan, OpenCL Adreno), and Android (arm64). Some build variants (KleidiAI, OpenEuler) remain disabled. The update is signed with GitHub’s verified GPG key, ensuring integrity. This incremental but practical feature lowers friction for deploying llama.cpp as a local AI backend.
- Adds --cors-allow-origin, --cors-allow-methods, etc. for cross-origin requests
- Pre-built binaries for macOS, Linux, Windows, Android across CPU/GPU backends
- Enables web apps on other domains to securely call local llama.cpp servers
Why It Matters
Makes local LLM servers compatible with modern web apps, simplifying developer workflows for AI prototyping.