OpenViSTAI · Agents · Blockchain · Crypto
Submit a project
Menu
AI Infrastructure & Serving

AI Grid

AI Grid is self-hosted software that connects computers into a shared infrastructure for AI inference and training. It provides a common endpoint that routes requests to computers serving the requested model, and includes command-line workflows for joining model-serving machines and using models.

Explore the practical guide ↓

Independent OpenViST project profile, based on linked sources.

AI Grid — OpenViST project profile in AI Infrastructure & Serving.
OpenViST project illustration. See the linked sources for product details.
Understand the project

Practical overview

When it fits

Use AI Grid when you want a shared endpoint for models running across computers you control, or want to explore the repository's documented media and experimental training workflows.

Who it is for

  • People pooling computers they already own for local AI inference.
  • Developers using OpenAI-compatible applications who want those apps to reach models on a shared grid.
  • Users exploring distributed model fine-tuning on their own hardware.

Documented uses

  • Route application requests to the computer serving a requested model.
  • Join a computer running a model to a grid and inspect the grid's available models.
  • Use command-line prompts with a joined model.
  • Configure OpenAI-compatible apps to send requests to the grid endpoint.
  • Set up documented ComfyUI workflows for image generation, image editing and image-to-video.
Before you choose

Requirements & limits

The repository describes the local grid as requiring no account or sign-in. The CLI installation instructions name macOS and Linux.

  • The grid host routes requests and does not need a graphics card; model-serving machines run the inference engine.
  • Grid and engine are separate roles: Grid coordinates access while engines such as llama.cpp or ComfyUI serve models or media tasks.
  • The documented quickstart uses GGUF models and can require multi-gigabyte downloads.
  • The README marks training as experimental; inference setup is documented separately.
  • The README directs users with computers in different locations to a remote mode, separate from the local quickstart.

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

Set up a two-computer inference grid

Install the CLI, connect a model-serving computer to a grid, and send it a prompt.

Before you start

  • Computers running macOS or Linux with network access between them; the README also allows the host and model runner to be the same computer.
  • A GGUF model and an engine such as llama.cpp, installed and pulled using the documented Grid commands.

Steps

  1. Install the Grid CLI on each computer using the project's documented installation command.

  2. On the routing host, run grid start and note the printed grid address.

  3. On a model-serving computer, install an engine, pull a model, then run grid join with the grid address, model filename, advertised model name and computer label.

  4. On the host, use grid models to inspect available models and grid chat to send a prompt; OpenAI-compatible apps can use the endpoint and environment values printed by grid info --env.

Expected outcome

The model-serving computer is joined to the grid, and the host can route a prompt to its advertised model; compatible apps can be pointed at the grid endpoint.

Guide limitations

  • The documented install command covers macOS and Linux.
  • Model downloads may be several gigabytes.
  • This recipe describes inference setup, not the separately documented experimental training workflow.
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
AI GridRoute application requests to the computer serving a requested model.Self-hosted, ServerThe repository describes the local grid as requiring no account or sign-in. The CLI installation instructions name macOS and Linux.
MorphingdbImport a libtorch model and create predictions from values in a PostgreSQL table.Docker, Linux, Self-hosted, ServerThe README documents a Docker image and a build-from-repository option, followed by connecting to PostgreSQL and enabling the pgdl extension.
Gateway — pyadavSend chat-completion requests to a provider through a common server endpoint.ServerThe repository is public and identifies the project as Apache-2.0 licensed. README installation instructions include binaries, platform package managers and Docker.
Evidence and transparency

Sources & review

Official repository ↗
Source record details
Source availability
Active
Sources checked
14 October 2026
Original website title
autonomous-ai/autonomous-grid · Official homepage title · retained in its source language
How we review sourcesSuggest a correction