Hyperspace (P2P inference network)

Decentralized peer-to-peer AI inference network. 2.7M+ CLI downloads, 2M+ active nodes globally as of April 2026. Three-tier model routing (local registry → DHT → gossip broadcast) supports any GGUF model. The April 2026 milestone: 32 anonymous nodes collaboratively trained a language model in 24 hours — the first cross-consumer-device training run with no trusted infrastructure.

By Fredoline Eruo·Last verified May 6, 2026·12,000 GitHub stars

Overview

Pros

True P2P inference — no centralized server dependency
Three-tier model routing finds any node with the model loaded
Browser client (WebLLM) plus CLI plus tray app
Cache layer eliminates redundant computation across the network

Cons

Inference latency depends on network mesh state
Privacy model still maturing — verify before sending sensitive data
Smaller model selection vs running locally with full Ollama catalog

Compatibility

Operating systems	macOS Linux Windows Browser
GPU backends	consumer GPUs via node-llama-cpp Apple Silicon WebLLM in browser
License	Open source · free (OSS) — pay-per-block on the live network

Get Hyperspace (P2P inference network)

Official site

https://hyper.space

GitHub

https://github.com/hyperspaceai/aios-cli

Frequently asked

Is Hyperspace (P2P inference network) free?

Hyperspace (P2P inference network) has a paid tier (free (OSS) — pay-per-block on the live network). Check the pricing page for current terms.

What operating systems does Hyperspace (P2P inference network) support?

Hyperspace (P2P inference network) supports macOS, Linux, Windows, Browser.

Which GPUs work with Hyperspace (P2P inference network)?

Hyperspace (P2P inference network) supports consumer GPUs via node-llama-cpp, Apple Silicon, WebLLM in browser. CPU-only inference is also possible but slow.

Reviewed by RunLocalAI Editorial. See our editorial policy for how we evaluate tools.