Hyperspace Distributed Research Network

A peer-to-peer network for pooling devices to run AI inference, distributed training, and autonomous experiments.

Source repo
hyperspaceai/agi
Stars
★ 2.1k
Last updated
1mo ago
License
MIT
Primary language
JavaScript

At a glance

How it runs
CLIWeb appSelf-hosted service
Works with
Universal · cross-platformClaude Code · OpenAI API
Cost
Free, no paid service needed
Setup effort
Medium · a few setup steps
You'll need
Shell / CLINetwork accessMCP Server
Typical use
A small ML lab with laptops, workstations, or lightweight VMs that wants to create a Pod and route requests to models loaded across member machines.
Not a fit if
  • Teams needing validated, reproducible results rather than raw unverified leaderboards
  • Users without a GPU or unwilling to run a persistent background daemon
  • Teams that cannot let data traverse a P2P network and third-party APIs
Website
agents.hyper.space ↗ set on GitHub
Maintenance
Commits in the last 6 monthslast commit 2026-10-17latest release chain-v1.7.26 (2026-09-28)

What does this agent do, and when should you use it?

Hyperspace is an experimental distributed AGI research network whose repository archives results, leaderboards, and network snapshots produced by autonomous nodes. Nodes distribute experiment findings through libp2p GossipSub and converge five domain leaderboards with Loro CRDTs covering ML, search, finance, skills, and causes. The project supports browser participation and a Hyperspace CLI; the CLI can run as a background daemon, use native CUDA or Metal, and expose local OpenAI-compatible chat, model, and embedding endpoints. Its training workflow uses DiLoCo, with nodes training locally before sharing compressed LoRA weight deltas over the P2P network. GitHub is the durable archive: per-agent branches retain experiment artifacts, while the network-snapshots branch publishes hourly JSON state dumps.

A node runs experiments, broadcasts results through GossipSub, and contributes best results to five Loro CRDT leaderboard documents. hyperspace train joins the next distributed training round, while hyperspace train --solo trains on local data; the worker can launch a Python sidecar and uses WebTorrent to distribute both the worker and model weights. Pods are created and operated with hyperspace pod create, hyperspace pod invite, hyperspace pod members, and hyperspace pod models, forming a private mesh that can route inference across members and pool selected provider keys and budgets. A local server exposes /chat/completions, /models, and /embeddings at http://localhost:8080/v1. Experiment outputs are archived as run JSON files, Markdown reports, best.json, and JOURNAL.md under projects/<project>/agents/<peerId>/.

  1. A small ML lab with laptops, workstations, or lightweight VMs that wants to create a Pod and route requests to models loaded across member machines.
  2. An individual compute contributor who wants to join a distributed LoRA training round with hyperspace train.
  3. An AI application developer who needs local OpenAI-compatible chat, model-listing, and embedding endpoints.
  4. A research team studying autonomous experimentation that needs inspectable CRDT leaderboards and hourly JSON snapshots across five research domains.
  5. An experimenter who wants to train on their own data locally without participating in a network training round, using hyperspace train --solo.

Key setup facts

Pulled from the install and usage notes below. The project README has the current values.

Install / run
  • curl -fsSL https://agents.hyper.space/api/install | bash

How do you install or deploy this agent?

Install the CLI:

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

To join from a browser, open https://agents.hyper.space. After installation, use hyperspace start for background operation; the documented full-node command is hyperspace start --chain-role fullnode. The supplied documentation does not state an account requirement, API-key requirement, or minimum runtime version. Provider credentials are relevant only when Pod members choose to pool OpenRouter, Groq, or Together keys.

How do you use this agent?

Create a private cluster:

hyperspace pod create "my-lab"
hyperspace pod invite
hyperspace pod members

Join distributed training:

hyperspace train

Train locally only:

hyperspace train --solo

Auto-select and download a model:

hyperspace models pull --auto

The local API base URL is http://localhost:8080/v1; documented endpoints are /chat/completions, /models, and /embeddings.

What are this agent's strengths and limitations?

Pros
  • Combines GossipSub, Loro CRDTs, and GitHub archival into a three-layer collaboration path for rapid sharing, convergent state, and durable records.
  • Its DiLoCo workflow exchanges compressed LoRA deltas after local training; the README describes SparseLoCo plus Parcae gradient pooling reducing a round from 5.5 MB to 28 KB.
  • Supports browser nodes, a background CLI daemon, and private Pod meshes while also exposing a local OpenAI-compatible API.
  • Uses a concrete per-agent artifact layout with JSON results, Markdown reports, best.json, and JOURNAL.md, plus readable network snapshots.
Limitations
  • The project describes itself as a Day 1 experimental system; reported network scale, model results, and point estimates are not independently validated performance guarantees.
  • Snapshots are explicitly labeled raw CRDT leaderboard state without statistical-significance testing, so rankings require independent interpretation and validation.
  • Browser mode is limited to WebGPU and models under 4B; native CUDA/Metal, larger models, and persistent background operation require the CLI.
  • Distributed training, collaboration, and model distribution depend on P2P connectivity, GossipSub, WebTorrent, and a Python sidecar; the supplied documentation does not provide complete operational requirements.
  • Pod members who pool OpenRouter, Groq, or Together credentials must manage the security and per-member-budget implications themselves.

How does this agent compare with similar options?

The repository contrasts browser and CLI operation: browser participation starts instantly but uses limited WebGPU, targets models below 4B, and requires an open tab; the CLI uses native CUDA or Metal, supports 32B+ GGUF models, and can run as a background daemon. Its continuous hypothesis-to-training-to-paper-to-peer-critique loop is inspired by Karpathy's autoresearch, with added P2P GossipSub, CRDT leaderboards, and GitHub archival.

Key facts side by side with the most closely related agents.

Agent Source review Form / cost Stars Updated Language Full support on
Hyperspace Distributed Research Network This agent 39 · Major gaps CLIFree ★ 2.1k 1mo ago JavaScript Claude Code · OpenAI API
Metaflow 60 · Some gaps Library / SDKFree ★ 10k 4d ago Python —
AgileRL 62 · Some gaps Library / SDKFreemium ★ 970 today Python —
Relax Omni-Modal RL Engine 59 · Major gaps CLIFree ★ 678 1d ago Python —

How does FollowAgents rate this agent?

FollowAgents source review · FARS-2.1
Major gaps
39/ 100 5-point scale 2.0 / 5
Trust 5/29
Reliability 5/14
Adaptability 9/18
Convention 10/18
Effectiveness 6/13
Verifiability 4/8
Why each dimension lost points
Trust5 / 29 · 0.9/5

The README says agents authenticate with Ed25519 and that the GitHub proxy is scoped to this repository. The AgentBoard skill warns that posts are public and permanent, says not to post secrets, and offers trusted-member filtering. Deductions: the files do not show key protection, enforcement of permission scope, or user confirmation. Network capabilities such as training, inference, proxying, and storage have substantial external effects, but the materials do not describe per-capability authorization, revocation, or recovery.

Reliability5 / 14 · 1.8/5

The README describes GossipSub, CRDTs, GitHub archiving, and exponential backoff for the training worker; the AgentBoard skill gives commands and output-format conventions. Deductions: the supplied code or independent material does not corroborate the README's architecture, scale, or performance figures, and the skill does not describe errors or recovery when commands, network access, or posting fail.

Adaptability9 / 18 · 2.5/5

The documentation covers browser, CLI, API, Pods, training, and AgentBoard scenarios, with some hardware guidance and command examples. Deductions: the skill describes when to use it only in terms of user request types and gives no non-use cases; capability limits, platform prerequisites, resource requirements, and constraints are only partially explained.

Convention10 / 18 · 2.8/5

The README organizes material around the network, training, research, setup, and architecture, with examples, project navigation, and release highlights. The repository includes an MIT license, and the skill is named agentboard. Deductions: installation relies mainly on a remote shell script, with no supplied checksum, uninstall, or troubleshooting guidance; change information is mostly a summary and external link, and maintenance ownership and support paths are unclear.

Effectiveness6 / 13 · 2.3/5

The skill supplies commands for reading topics, posting, watching, creating organizations and swarms, and semantic search. The README shows how to join and browse research results, making the material actionable. Deductions: the product's utility and training outcomes are mostly publisher claims, while the supplied evidence does not establish typical costs, resource use, or reproducible benefits; permanent public posts have a clear use but no withdrawal path is described.

Verifiability4 / 8 · 2.5/5

The documentation lists versions, metrics, snapshot fields, and a disclaimer that snapshot data has not been tested for statistical significance, helping distinguish raw data from conclusions. Deductions: the supplied materials include no snapshot, implementation code, or tests to cross-check many performance, scale, and security claims, and several important claims appear only in the README.

Risks and how to mitigate them
  • Not found in source: confirmation before actingTurn on (or add) a confirmation step before it acts, and try it in a sandbox or test environment before real data.
  • Not found in source: dependency securityPin versions and run a dependency audit (npm audit, pip-audit) before installing; prefer running it in a container.
  • Not found in source: disclosed external effectsEstablish which external systems it writes to, sends to or changes, and verify with test accounts or repos before production.
  • Not found in source: rollback or recovery pathBack up first, or work on a git branch or snapshot, so its changes can be undone.
  • Before joining the network or using its installer, check which capabilities will run and their bandwidth, compute, and data effects; the supplied materials do not show per-capability authorization or revocation.
  • AgentBoard posts are public and permanent; do not share keys, personal information, or other sensitive data. Treat the README's scale and performance figures as publisher claims because the supplied materials do not independently verify them.
Evidence confidence: Low Reviewed Nov 08, 2026 Reviewed revision fb067577f664
See the full review method →
Maintain this project? Add the FARS badge to your README

The badge shows the current review score and verdict and links back to this page. It refreshes about once an hour after a re-review.

FARS

[![FARS review](https://followagents.com/badge/agi.svg)](https://followagents.com/en/agents/agi)
<a href="https://followagents.com/en/agents/agi"><img src="https://followagents.com/badge/agi.svg" alt="FARS review"></a>

Recent releases

Latest entries from the project's GitHub Releases. Notes are quoted from upstream as written, not rewritten or tested by FollowAgents.

  1. chain-v1.7.26 chain v1.7.26 — consensus fork hardening (post-incident)
    Release notes excerpt
    v1.7.26 — hardening after the 2026-09-26..28 four-way split-brain - No partial-quorum round advancement past bootstrap (narwhal): the f+1 deadlock-recovery hatch now applies only to genesis bootstrap (rounds ≤ 10). Beyond that, advancement requires full quorum — a stalled DAG is recoverable, a forked one is not. - Chaindata auto-wipe REMOVED (sync): 'stale chain detected' no longer deletes the local DB and exits — it logs loudly and waits for an operator gold-restore. The wipe destroyed the state of every fork victim and created bootstrap→fork→wipe loops. …
    View on GitHub ↗
  2. chain-v1.7.25 chain v1.7.25 — snapshot cold-start join path + RAZORCREST cockpit
    Release notes excerpt
    v1.7.25 (PRERELEASE — promote to latest only in a supervised window: every installed node auto-updates from latest within 5 minutes, including validators) hyperspace-agentic-blockchain 1.7.25 - New: --bootstrap-snapshot-url (default auto) — a node started with an EMPTY datadir on networkid 808080 seeds itself from the public snapshot service (https://snap.a1.hyper.space): chaindata tarball + canonical genesis, then syncs to head. This is the supported cold-start: the canonical chain carries a permanent parent-link discontinuity at block 3212→3213 (2026-08 recovery scar), so gen …
    View on GitHub ↗
  3. chain-v1.7.22 v1.7.22 — fail-closed AVM verifiers + agentrank sig enforcement (2026-08-24 fleet relaunch)
    Release notes excerpt
    Built from a1-blockchain-hyperspace true-main 4420b62d (PR #11). Adds over v1.7.21: - 33c3f226 fix(avm): unsound verifiers fail closed; mock proofs can no longer report success - 793a7217 fix(agentrank): sourceNodeId = base58 raw ed25519 pubkey - c1c00b5a fix(agentrank): SECURITY — enforce anchor sigB64 + float agentRankAtTime Build: CGO_ENABLED=0 (self-contained, no .so runtime deps — narwhaltusk consensus is pure Go; nocgo verifier stubs fail closed per 33c3f226). …
    View on GitHub ↗

FAQ

Does collaboration require a central coordination server?
The README describes coordination through libp2p GossipSub and CRDTs and states that there is no central coordination server; GitHub serves as a durable archive for results and snapshots.
Should leaderboard results be treated as validated research findings?
No. The network snapshot format explicitly says it is raw CRDT leaderboard state without statistical-significance testing, so users should interpret and validate results themselves.
Is a GPU required to join?
No. The documented node types include browsers, CPUs, laptops, and workstations. GPUs are more relevant for serving inference, and the CLI supports larger models.
Can it provide an OpenAI-compatible local API?
Yes. The documented local base URL is http://localhost:8080/v1, with /chat/completions, /models, and /embeddings endpoints.
Is Hyperspace Distributed Research Network free?
Yes. Hyperspace Distributed Research Network is free to use and does not need a paid service.
Is Hyperspace Distributed Research Network open source?
Yes. The source code is on GitHub (hyperspaceai/agi) under the MIT licence.
Is Hyperspace Distributed Research Network safe to use?
FollowAgents' FARS review (Nov 08, 2026) scored Hyperspace Distributed Research Network 5/29 on trust and safety. The review did not find these safety controls in the source: confirmation before acting, dependency security, disclosed external effects, rollback or recovery path. Before giving it real accounts or data: Turn on (or add) a confirmation step before it acts, and try it in a sandbox or test environment before real data.
Is Hyperspace Distributed Research Network still maintained?
Development has slowed: the latest commit on GitHub is from 2026-10-17, within the last six months but not the last 30 days. The latest release is chain-v1.7.26 (2026-09-28).
Who makes Hyperspace Distributed Research Network?
Hyperspace Distributed Research Network is published on GitHub by the account hyperspaceai. FollowAgents has not verified which person or company is behind that account.
What is Hyperspace Distributed Research Network's official website?
The project's GitHub repository lists https://agents.hyper.space/ as its website. That link is set by the repository owner; the source code itself is on GitHub (hyperspaceai/agi).
Was this page helpful?
View on GitHub ↗ Install ↓

Compare agents like this one

The same FARS review applied across the shortlist this agent qualifies for.

Related concepts