Turn distributed GPUs into your next AI breakthrough. Run inference, find compute, and follow every execution from request to receipt.
Start in the Playground. Connect through an API. Put your own GPU to work. There’s a path for every side of the network.
Try AI models in a focused Playground. See where your response ran and inspect the job behind it.
Open PlaygroundCompare GPU memory, supported models, regions, and hourly rates. Fund a rental with USDG escrow.
Explore GPUsConnect a machine with the worker agent and Ollama. Publish your capabilities and accept network inference.
Become a providerPermute tries eligible network workers first and uses cloud APIs when needed. Execution source is disclosed, so every result comes with context.
Explore the networkUse an OpenAI-compatible chat endpoint. Create a key, send your first request, and inspect routing information in the response.
Build with the APIconst response = await fetch(
"https://permutecompute.xyz/api/chat/completions",
{
method: "POST",
headers: {
Authorization: "Bearer YOUR_API_KEY",
"Content-Type": "application/json"
},
body: JSON.stringify({
model: "qwen-lite",
messages: [{ role: "user", content: "Hello, Permute" }]
})
}
);