The people's inference layer

Your Mac is a money machine. You just don't know it yet.

NodeHive is a marketplace where anyone can list their hardware as an AI inference endpoint. We orchestrate requests. You earn from idle compute. Consumers get AI at a fraction of cloud prices.

40% of global GPU capacity sits idle
10x cheaper than cloud AI providers
70B+ parameter models on a Mac Mini
How it works

Three steps. Zero complexity.

01

List your endpoint

Point NodeHive at your running model. Mac Mini, GPU rig, whatever. We verify it's live and add it to the network.

02

We route the traffic

Our orchestration layer balances requests across all listed endpoints. OpenAI-compatible API. One line to switch any app over.

03

Everyone earns

Suppliers get paid per request. Consumers get inference at basement prices. Early listers keep 100% of proceeds for life.

For Suppliers

Turn idle hardware into passive income

  • ✓List any Mac with Apple Silicon, or any NVIDIA GPU
  • ✓Choose which models you serve
  • ✓Early listers get 100% of proceeds, forever
  • ✓Share your endpoint link to bring your own customers

For Consumers

AI inference without the cloud markup

  • ✓OpenAI-compatible API, switch with one line of code
  • ✓Access Llama, Mistral, DeepSeek, Qwen and more
  • ✓Prices start free, paid tiers are 10x cheaper than cloud
  • ✓Requests routed to the fastest available endpoint
Developer friendly

One line to switch.

If you already use the OpenAI SDK, you're already compatible.

// Before: paying cloud prices
const client = new OpenAI({ baseURL: "https://api.openai.com/v1" })

// After: 10x cheaper, same API
const client = new OpenAI({ baseURL: "https://api.nodehive.dev/v1" })

const res = await client.chat.completions.create({
  model: "llama-3-70b",
  messages: [{ role: "user", content: "Hello" }]
})

The world's compute is already here. It just needs a marketplace.

Millions of Macs and GPUs run idle every day. NodeHive connects them to the people and companies that need inference, at prices that make cloud providers irrelevant. The infrastructure revolution isn't about building more data centers. It's about using what already exists.