Gophercamp2026
Back to all sessions

Kronk: Hardware accelerated local inference

by William Kennedy

About this Session

In this talk Bill will introduce Kronk, a new SDK that allows you to write AI based apps without the need of a model server. If you have Apple Metal (Mac), CUDA (NVIDIA), or Vulkan, Kronk can tap into that GPU power instead of grinding through the work on the CPU alone. To dog food the SDK, Bill wrote a Model Server that is optimized to run your local AI workloads with performance in mind. During the talk, Bill will show how you can use Agents like Cline and Kilo Code to run local agentic workloads to perform basic work.