OpenViSTAI · Agents · Blockchain · Crypto
Submit a project
Menu
Local & Self-hosted AI

TensorFold

TensorFold is an application for running open language models locally on Apple Silicon Macs and NVIDIA GPUs, using native Metal or CUDA inference. Its macOS app bundles the TensorFold engine, and its guide also documents a separate command-line engine for serving models and connecting compatible clients.

Explore the practical guide ↓

Independent OpenViST project profile, based on linked sources.

TensorFold — OpenViST project profile in Local & Self-hosted AI.
OpenViST project illustration. See the linked sources for product details.
Understand the project

Practical overview

When it fits

Use TensorFold when you want to run an open model on your own supported hardware, either through the Mac app or the separately documented command-line server workflow.

Who it is for

  • Mac users who want to download and chat with a compatible local model.
  • NVIDIA GPU owners seeking local language-model inference.
  • Developers who want compatible OpenAI- or Anthropic-style clients to use local inference.

Documented uses

  • Run an open language model on an Apple Silicon Mac or NVIDIA GPU.
  • Download and load a compatible model for local chat in the Mac app.
  • Serve local model inference for compatible OpenAI and Anthropic clients.
Before you choose

Requirements & limits

The documented Mac app requires an Apple silicon Mac running macOS 13.3 or later. The homepage identifies TensorFold as Apache-2.0 licensed.

  • The documented app path is for Apple silicon Macs on macOS 13.3 or later; the command-line guide is a separate installation path.
  • A model must be compatible with the hardware checks, and model files require disk space.
  • The guide lists OpenAI- and Anthropic-client connection workflows for local inference.
  • The homepage identifies Metal and CUDA as native inference paths for Apple Silicon and NVIDIA GPUs.

Check current requirements in the official documentation. Source-based guidance does not mean OpenViST has run or independently tested this project.

Source-based practical guide

Install the TensorFold Mac app and start a local chat

Set up the macOS app, load a compatible model and send a first message.

Before you start

  • Apple silicon Mac running macOS 13.3 or later
  • Enough disk space for the app and selected model
  • A compatible model that passes the hardware checks

Steps

  1. Download the Mac app, open the DMG, drag TensorFold into Applications and launch it.

  2. Choose Minimal for a local model server or Full for the chat and work interface; select the built-in TensorFold engine during setup.

  3. In the app’s Models area, choose a compatible model, download it, load it and wait for its state to show Running.

  4. In Full mode, open Chat and send a message.

Expected outcome

A compatible local model is loaded in TensorFold and can respond in a chat.

Guide limitations

  • The documented app path requires Apple silicon and macOS 13.3 or later.
  • Model loading depends on the model passing hardware checks and having sufficient disk space.
  • The guide says app setup does not automatically download, load or send a message.
Documentation 1 ↗
COMPARE OPTIONS

Compare options for Model inference and serving

These projects document a common task and share a project role. This is a source-based comparison, not an execution benchmark or a guarantee of interchangeability.

ProjectDocumented purposePlatformAccess and requirements
TensorFoldRun an open language model on an Apple Silicon Mac or NVIDIA GPU.macOS, Self-hostedThe documented Mac app requires an Apple silicon Mac running macOS 13.3 or later. The homepage identifies TensorFold as Apache-2.0 licensed.
TurboLLMStart a local model runner and chat with models through a browser interface.Windows, macOS, Linux, Android, Desktop, Web, Self-hostedThe site says TurboLLM runs locally with no account and works offline. Desktop setup requires Node.js 22 or newer; Android is documented as a standalone app without Node.js or a terminal.
Bodhi AppDiscover and download GGUF models from Hugging Face that suit local hardware.Linux, macOS, Self-hosted, WindowsThe README documents a Mac desktop release and Docker images for CPU, CUDA, ROCm and Vulkan configurations. The project identifies itself as open source.
Evidence and transparency

Sources & review

tensorfold.dev ↗Guide documentation ↗
Source record details
Source availability
Active
Sources checked
3 November 2026
Original website title
TensorFold | Run LLMs, Fast. · Official homepage title · retained in its source language
How we review sourcesSuggest a correction

Related options

StarLight — A separate TensorFold enterprise control plane for managing organizational model fleets, identity, routing and usage controls; it is distinct from the local inference app.