Your app asks.
p0wer finds the GPU.
Build on p0wer with one familiar API. Your app sends the request and the network finds the available compute.
Change two lines.
Connect MetaMask once in Dashboard, create a wallet-linked API key, and point any OpenAI client at `p0wer-free`.
client = OpenAI(
base_url="https://YOUR-DOMAIN/api/v1",
api_key="YOUR_API_KEY"
)
r = client.chat.completions.create(
model="p0wer-free",
messages=[{"role":"user","content":"Explain p0wer."}],
stream=True
)
Simple on the outside.
Streaming SSE, OpenAI-shaped chat completions, model availability, and capacity that reflects actual online workers.
Start free. Grow with the network.
The browser lane is the easiest to populate. Native and swarm lanes only report available when those worker types are really connected.
p0wer-free
Browser-worker lane for the free launch experience.
p0wer-native
Optional higher-power native lane for dedicated workers.
p0wer-swarm
Distributed large-model lane that stays offline until a real swarm is serving.
No GPU required to call it.
API users consume network compute. Contributors are a separate supply side.
Full docs60 requests / hour
Default per wallet-linked key. The launch API does not require you to own or contribute a GPU.