Hyperspace AGI

Overview

Hyperspace AGI is the first distributed autoresearch system — a living research repository where thousands of autonomous AI agents collaboratively run experiments, share findings via peer-to-peer gossip, and push results to GitHub.

Built on top of the Hyperspace decentralized inference network (2M+ nodes, 3.6M+ downloads, libp2p/IPFS protocol stack), it extends the Karpathy autoresearch pattern from single-agent/single-GPU to a massively distributed multi-agent system.

The repo itself is a research artifact — agents push experiment results to per-agent branches on GitHub, and a network node publishes consolidated snapshots (snapshots/latest.json) every hour to the main branch. No narrative, no curation — raw CRDT leaderboard state from the live network.

Key stats (as of June 11, 2026): 1,923 stars (up from 696 in mid-March), 231 forks, MIT licensed, created March 8, 2026 (2 days after Karpathy's autoresearch). The README reports 660 autonomous agents have run 27,247 experiments across the system.

An honest aliveness caveat: the consolidated hourly snapshots and main-branch commits stopped on April 17, 2026 — snapshots/latest.json has not been updated in nearly two months. Agent branches are still pushing (the repo's last push was June 11, 2026), so the research loop is running, but the curated archive layer has stalled.

---

Architecture

The Three-Layer Collaboration Stack

Hyperspace AGI's key innovation is its coordination architecture. Every research domain uses three layers, each at different latency:

1. GossipSub (~1 second) — Agent finishes an experiment, broadcasts the result to all peers instantly via libp2p GossipSub
2. CRDT Leaderboard (~2 minutes) — Loro conflict-free replicated data types sync each peer's best result. New nodes read the full leaderboard on connect — zero cold start
3. GitHub Archive (~5 minutes) — Best results pushed to hyperspaceai/agi per-agent branches. Permanent, human-readable record

This is the most sophisticated multi-agent research coordination system in the open-source ecosystem. The layered approach solves a real problem: real-time inspiration (gossip) for fast iteration, convergent state (CRDT) for consistency, and durable archival (git) for reproducibility.

Research Domains

Agents operate across 5 domains simultaneously, each with its own metric and CRDT leaderboard:

This multi-domain approach is a significant departure from other autoresearch tools, which are single-domain. Hyperspace AGI treats research as a portfolio problem — agents can specialize or generalize.

The Research Loop

Each agent runs a continuous cycle inspired by Karpathy's autoresearch:

1. Hypothesize — Generate ideas: "What if we use RMSNorm instead of LayerNorm?"
2. Experiment — Run on whatever hardware is available (browser tab to H100)
3. Share — Broadcast results via P2P gossip
4. Synthesize — Accumulate enough experiments, write a research paper
5. Peer Review — Other agents read, critique, and score papers 1-10
6. Evolve — Papers scoring 8+ are flagged as breakthroughs, feed back into Stage 1

Multiple agents can train the same model collaboratively via DiLoCo — each trains locally for H steps, then shares compressed weight deltas. Automatic fallback to solo training if no peers are available.

---

Network Infrastructure

The Hyperspace Node Network

The AGI repo is the research layer built on top of a much larger infrastructure play:

• 2M+ active agents across the P2P network
• 3.6M+ downloads of the Hyperspace node client
• 6 bootstrap nodes across US, EU, Asia, South America, and Oceania
• Built on libp2p (same protocol as IPFS)
• OpenAI-compatible local API at localhost:8080/v1 — any tool that speaks OpenAI can use Hyperspace as a backend

Node Capabilities

Each node can run any combination of 9 capabilities:

Points Economy

Two earning streams incentivize participation:

Presence points (pulse rounds every ~90s):

• Base 10 points per epoch
• Uptime bonus: U(t) = 1 + 0.2 * ln(1 + t/12) — 30-day nodes earn 83% more
• Capability bonus: more capabilities = more points

Work points (task receipts):

• tokens * cost_per_token * model_multiplier * uptime_bonus

The points system creates a SETI@home-style incentive structure — contribute compute, earn rewards. The research capability (+12%) is the highest-weighted, signaling the network's priority on autoresearch.

Pricing

There are no paid tiers as of June 2026 — joining as a browser or CLI agent is free, and participation is compensated in points rather than charged. Pods (shared agent groups) use bring-your-own provider keys (OpenRouter, Groq, Together among 27+ supported providers) with optional budget controls, so any LLM spend is pass-through to your own provider accounts. Points remain non-tradeable, but the network is listed on airdrop trackers in anticipation of a token conversion.

---

What Changed Since March 2026

The project shipped aggressively between March and mid-April, then went quiet on the main branch:

• Distributed training went live — a 32-node, 24-hour DiLoCo run with 195× gradient compression, moving collaborative multi-agent training from roadmap to demonstrated capability
• Hyperspace A1 blockchain — a chain with Mysticeti consensus was integrated, with 54 chain releases shipped (latest chain-v1.7.8, April 29, 2026)
• CLI velocity — 96 CLI releases from v5.0.0 to v5.20.0
• Claude Code MCP integration — CLI v5.19.0 (April 14, 2026) exposed the full Hyperspace CLI as 82 MCP tools for Claude Code, spanning node, models, inference, pods, providers, budgets, and sandboxes
• Pods and budgets — resource pooling with shared provider keys, four budget modes, and encrypted portable "Pod Capsules"
• Snapshot stall — hourly consolidated snapshots to main stopped April 17, 2026, even as agent branches keep pushing

The blockchain integration confirms the crypto trajectory the March profile flagged as a watch item.

---

Competitive Position

Strengths

• Network effects at scale — 2M+ nodes is a massive moat. No other autoresearch system has this kind of distributed compute
• Multi-domain research — 5 simultaneous research domains vs single-domain competitors
• Production P2P infrastructure — Real libp2p network with DHT, gossip, NAT traversal. Not a prototype
• Zero-install browser option — Lowest barrier to entry in the category
• Hourly GitHub snapshots — Full transparency, anyone can analyze the raw CRDT state

Weaknesses

• Token/crypto adjacent — now confirmed — The A1 blockchain (Mysticeti consensus) shipped in spring 2026, and the network is listed on airdrop trackers. The "earn while you compute" framing increasingly attracts node farmers over researchers
• Research quality unproven — 1,923 stars vs Karpathy's 86k. No published breakthrough results yet
• Archive layer stalled — the hourly GitHub snapshot pipeline, a headline transparency feature, has not published since April 17, 2026
• Complexity — 9 capability types, 5 research domains, 3 coordination layers, 7-step verification protocol. A lot of moving parts
• Agent autonomy unclear — How much genuine research insight vs mechanical parameter sweeps?

vs. Karpathy's autoresearch

vs. autoresearch-at-home

Hyperspace AGI is what autoresearch-at-home aspires to be — a SETI@home for AI research — but with a production network already running. autoresearch-at-home has 188 stars and is still conceptual; Hyperspace has 2M nodes and live research loops.

---

What Developers Say

No substantive verbatim developer commentary on Hyperspace AGI surfaced in searches of Hacker News and Reddit as of June 2026 — the Hacker News threads matching "Hyperspace" concern an unrelated macOS utility, and the discussion that does exist around the network centers on points farming and airdrop speculation rather than research quality. Star growth (696 → 1,923 in three months) and continued agent-branch pushes are currently the best available adoption signals.

---

What to Watch

• Research output quality — Can distributed agents produce genuine breakthroughs, or is this mostly parameter sweeping at scale? 27,247 experiments logged, zero published breakthroughs so far
• Token economics — The A1 blockchain is live and airdrop trackers are circling; when points convert to tradeable tokens, does it attract researchers or speculators?
• Snapshot pipeline recovery — whether the hourly consolidated archive (stalled since April 17) resumes is the clearest signal of whether the research layer is still a priority
• DiLoCo collaborative training — the 32-node, 24-hour run with 195× compression proved the mechanism; the question is whether it scales to useful models
• Domain expansion — 5 domains today, but the architecture supports arbitrary research targets
• Community research papers — The peer review loop (agents scoring each other's papers 8+) could produce interesting emergent research

---

Bottom Line

Hyperspace AGI remains the most ambitious entry in the autoresearch category — a distributed, multi-domain, multi-agent research system running on a 2M-node P2P network. Where Karpathy proved the pattern works for a single agent on a single GPU, Hyperspace is betting that intelligence compounds when thousands of agents share findings in real-time. The three-layer coordination architecture (gossip, CRDT, GitHub) is genuinely novel and solves real distributed systems problems.

Not recommended as a research platform to depend on: the curated archive layer has been stalled since April 17, 2026, no breakthrough results have emerged from 27,000+ experiments, and the shipped blockchain plus airdrop-tracker attention suggest the points economy is attracting node farmers faster than researchers. Recommended as the reference architecture for multi-agent research coordination — and the Claude Code MCP integration (82 tools) makes the network trivially explorable from a coding agent. Outlook: with Karpathy's repo dormant, Hyperspace is the only autoresearch system still running at scale; whether it stays a research project or completes the pivot to a crypto compute network will be decided by where the token lands.

---

Research by Ry Walker Research