logoGPU/docs
Dashboard← Back to site

API Overview

OpenAI-compatible inference API with live model availability and on-chain settlement.

Base URL

text
https://gpunetwork.xyz/v1

Endpoints

  • POST /v1/chat/completions — generate a chat completion (streamed or not)
  • GET /v1/models — list supported models and live ready-node counts
  • GET /v1/account/info — your account summary (key, usage, claimable)
  • GET /v1/network — public network stats (providers online, tasks completed)
  • GET /v1/providers — sanitised list of live providers
  • GET /v1/leaderboard — public Season 0 points leaderboard
  • GET /health — uptime + Redis check

Example

bash
curl https://gpunetwork.xyz/v1/chat/completions \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gpu-mesh",
    "messages": [
      {"role": "user", "content": "Explain decentralized inference in one sentence"}
    ],
    "stream": true
  }'

OpenAI SDK compatibility

Drop-in. Point any OpenAI client at our base URL and use your gpu_* key as the API key.

python
from openai import OpenAI

client = OpenAI(
    base_url="https://gpunetwork.xyz/v1",
    api_key="YOUR_API_KEY",
)

stream = client.chat.completions.create(
    model="gpu-mesh",
    stream=True,
    messages=[{"role": "user", "content": "hi"}],
)
for chunk in stream:
    print(chunk.choices[0].delta.content or "", end="", flush=True)
Rate limits
Per-key rate limits are enforced via the credit balance. Each request debits credits based on tokens and model multiplier; running out returns HTTP 402 insufficient_credits. New wallets get 10,000 starting credits; the dashboard offers daily faucet + points top-up.
Next →
Authentication