Cloudflare introduces a novel method of function calling that co-locates LLM inference with function execution, along with a new ai-utils package for enhancing the developer experience. This follows their mid-June announcement on traditional function calling, which allows users to leverage an LLM to generate structured outputs and pass them to an API call. The goal is to make building with AI as easy as possible by offering open-source model inference and a great developer experience. With embedded function calling and the ai-utils package, developers can create intelligent AI agents more easily, opening up endless possibilities.