Network

Workers and earning

Lend your GPU to the network and get paid per token you serve.

Anyone with a GPU can serve the network and get paid for every token they serve. There is no stake to put up and no datacenter required. Your node handles shares, not whole prompts.

Native worker (recommended)

Runs in the background through Ollama or vLLM. Serves the largest models and earns the top rate.

01
Install a runtime
Install Ollama or vLLM and make sure it can see your GPU. You also need Node 18 or newer.
02
Run the worker
Run the command below, replacing YOUR_TOKEN with the worker token for your account.

Terminal

npx @nuroaixyz/worker --token YOUR_TOKEN

Browser worker

Contribute straight from a browser tab with WebGPU, using Chrome or Edge. Nothing to install. It serves small, quantized models at the entry rate, and runs while the tab is open. Check your browser on the Earn page.

Native vs browser

	Native	Browser
Runtime	Ollama or vLLM	Chrome / Edge with WebGPU
Install	Node 18+, one command	None
Models	The biggest the network serves	Small, quantized
Runs	In the background	While the tab is open
Rate	Top rate / token	Entry rate / token

Getting paid

Every token you serve is metered and settled in $PRINET to your payout wallet. Users pay for inference; workers do not front anything.