Find tools › Open-source projects › Keep data private

TensorFold

Open-source project · by ashhart

Runs AI models directly on your graphics card [GPU] without requiring Python.

Some setupWell known

What it costs

Free and open source; you may pay for the AI model it uses.

Use it to

  • Serve open-weight AI models locally on Mac or Linux
  • Run a local AI server compatible with standard apps
  • Chat with local models via an optional desktop app

How to start

  1. Install TensorFold on macOS by running brew install ashhart/tensorfold/tensorfold.

  2. Download a model checkpoint using tensorfold pull mlx-community/Qwen3.5-2B-MLX-4bit.

  3. Start the server with tensorfold serve mlx-community/Qwen3.5-2B-MLX-4bit --name local-model.

Good to know. Requires Apple Silicon or an NVIDIA GPU with modern driver releases (CUDA 13+); experimental features like Living Weights can affect accuracy.

Before you install

  • Found automaticallyNot reviewed by a person yet. Check it before installing.
  • License: Apache-2.0You are allowed to use it under these terms.
  • Last updated 2026-11-18

More detail

TensorFold is a native language model engine written in Zig that serves open-weight models on Apple Silicon and NVIDIA GPUs.

Running AI models locally usually requires heavy Python environments and complex runtime setups. Reach for TensorFold when you want a single fast binary to serve open models locally with standard network interfaces.

Similar tools

Descriptions are written in plain English from each project's own documentation and refreshed by our agent; this one was last checked 2026-11-12. Spotted a mistake? · Open in Find tools · © 2026 Which AI Map