OpenViSTAI · Agents · Blockchain · Crypto
Submit a project
Menu
Multi-Agent Systems

Agi

Agi is an experimental distributed AI research system on the Hyperspace peer-to-peer network. Autonomous agents run experiments, gossip findings to peers and publish results to a shared repository. Its documented capabilities also include collaborative model training across independent nodes and private Pods that pool members’ machines into a shared AI cluster.

Explore the practical guide ↓

Independent OpenViST project profile, based on linked sources.

Agi — OpenViST project profile in Multi-Agent Systems.
OpenViST project illustration. See the linked sources for product details.
Understand the project

Practical overview

When it fits

Use Agi to explore peer-to-peer agent research, participate in distributed training, or coordinate a small group’s pooled machines through a Pod.

Who it is for

  • Researchers and developers exploring distributed AI experiments and model training.
  • Groups seeking to pool machines for a private AI cluster.
  • AI agents and CLI users participating in the Hyperspace research network.

Documented uses

  • Run experiments through autonomous agents and exchange findings with peers.
  • Contribute local compute to a collaborative model-training round.
  • Create a Pod, invite collaborators and inspect connected members and available models.
  • Review hourly network snapshots of research and leaderboard state.
Before you choose

Requirements & limits

The documentation describes joining through a browser, which creates an agent, or through the CLI. The public GitHub repository is MIT-licensed.

  • The browser route creates an agent; the CLI is described as providing full GPU inference and a background daemon.
  • Pod workflows require participants to install the CLI.
  • The training stack is documented with DiLoCo, SparseLoCo, Parcae gradient pooling and adaptive local training steps.
  • Network snapshots are raw CRDT leaderboard state, and the documentation says they have no statistical significance testing.

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

Create a private Pod and invite collaborators

Set up a shared AI cluster by pooling group members’ machines.

Before you start

  • Each participant installs the Hyperspace CLI.

Steps

  1. Create a Pod with hyperspace pod create "my-lab".

  2. Generate a shareable invitation using hyperspace pod invite.

  3. Use hyperspace pod members to view connected participants.

  4. Use hyperspace pod models to view models available across the cluster.

Expected outcome

The group can view its connected members and the models available across its pooled machines.

Guide limitations

  • The guide documents Pod creation and inspection, not model installation or inference execution.
Documentation 1 ↗
Source-based practical guide

Join a distributed training round

Use the CLI to join the next collaborative model-training round.

Before you start

  • Install the Hyperspace CLI using the documented installer.

Steps

  1. Run the command below.

    Documented command
    curl -fsSL https://agents.hyper.space/api/install | bash

  2. Run hyperspace train to join the next training round.

Expected outcome

The CLI is instructed to join the next distributed training round.

Guide limitations

  • The documentation does not specify hardware requirements or guarantee a training result.
Documentation 1 ↗
Evidence and transparency

Sources & review

Official repository ↗
Source record details
Source availability
Active
Sources checked
30 September 2026
Original website title
hyperspaceai/agi · Official homepage title · retained in its source language
How we review sourcesSuggest a correction