TensorFold Offers Native LLM Inference for Apple Silicon and NVIDIA GPUs
The ashhart/TensorFold project describes itself as a native LLM inference engine for Metal, CUDA and Vulkan. Its documentation says a Zig binary serves language models on Apple Silicon and NVIDIA GPUs by reading checkpoints, tokenizing requests and running Metal or CUDA kernels directly, without requiring Python or MLX. The supplied material contains no independent performance benchmarks, and the Vulkan implementation is not detailed beyond the project’s headline claim.
TensorFold could provide a more native deployment option for local language-model serving across Apple and NVIDIA hardware without a Python or MLX installation. The claims are documented by the project itself, however, rather than independently tested, and its Vulkan support remains unverified.
LLM Inference Engine for Metal, CUDA and Vulkan.

TensorFold
TensorFold serves language models from a Zig binary on Apple Silicon and NVIDIA GPUs. The engine reads checkpoints, tokenizes requests and runs Metal or CUDA kernels directly. Serving needs no Python or MLX installation.
Some source content is omitted. Continue reading on the original site.