Verified, then logged.
Sun 22 Nov 2026 · 00:15 KST / 15:15 UTC한국어 | ENSearch /

Breaking02 / IT

llama.cpp changes server tool-call controls

The release changes server tool controls and adds chat search, while its notes omit the licence and minimum VRAM requirements.

A developer whose face is out of frame checks external tool controls and grouped conversation search on a monitor beside a compact local server. AI-generated image

ggml-org announced llama.cpp v1.0.0 on 13 Nov, with changes to server tool controls and Windows model loading; its release notes do not specify a licence.

The notes cover changes since v0.9.0. They list a new unbuffered model-loading method for Windows, which loads model files without an intermediate buffer.

The server changes include fixes for flaws in request proxying and media URL handling. The notes also list a fix for handling cross-origin request settings, which control requests from other web origins.

The notes specify that tools must be explicitly enabled when the server is accessible beyond the local machine. External tool calls are cancelled when a client disconnects or a request times out.

The change list adds a chat-search tool and the ability to organise conversations into named groups.

The release notes do not specify the licence or a minimum graphics memory (VRAM) requirement.