⚡ Powered by Node.js + tfjs-node-gpu

Rent Massive GPU Clusters for AI Training — GEMM, Backprop & Tensor Ops at Scale

Node.js orchestration + tfjs-node / tfjs-node-gpu support. FinOps intelligence keeps costs low.

I'm an AI Engineer I'm a Hardware Provider $Nexus Wallet
Powered by decentralized providers PyTorch & TensorFlow ready Real-time monitoring
Renter Portal

For AI Engineers

Run compute on online GPUs and TPUs. Pay in $NEXUS per million input/output tokens when jobs settle.

  • Browse online provider hardware
  • $Nexus wallet for compute settlement
  • Pay only for tokens consumed
Browse Hardware Sign Up Free
Provider Portal

For Hardware Providers

List your GPUs, run the Node.js agent, and earn $NEXUS when engineers rent your machines.

  • List hardware — marketplace sets token rates
  • Agent heartbeats & job execution
  • Real-time earnings & uptime rewards
Start Listing Sign Up Free
Renter Portal

Train Faster. Pay Less.

Instant access to GPUs, TPUs, NPUs, FPGAs with high VRAM. Direct tfjs-node-gpu support for GEMM, backprop, and tensor ops.

🚀

Instant GPU Access

32–96GB+ VRAM nodes ready in seconds. Filter by TFLOPS, price, and backend.

⚙️

tfjs-node / tfjs-node-gpu

Matrix multiplication, vector ops, activations, transposition — routed to the right backend automatically.

📦

Pre-installed Environments

PyTorch, TensorFlow, and Node.js backends configured and ready for one-click deployment.

Live Network Overview

24-hour utilization and active training jobs from the marketplace

Live
Start Training Now — Filter by VRAM & TFLOPS
Renter Portal

Online Hardware

Browse online provider nodes and run compute via the renter SDK. You are billed via settlement only when jobs complete.

1. Sign In

Create an account or sign in to get your API credentials.

2. Install the Renter SDK

Programmatic access to browse hardware, submit jobs, and settle in $NEXUS.

3. Submit a Job

Pick a hardwareId from the grid below, choose a priority tier, and run compute — charges apply at settlement.

Run compute with the SDK

Sign in to personalize the example below.

Sign in to get your renter SDK example.

Full API docs · npm package

Provider Portal

Monetize Idle Hardware — Earn $NEXUS

List your GPUs and run the agent on your machine. Sign up to get started.

1. Sign In

One account to rent GPUs and list your own hardware.

2. List GPUs (Auto-Priced)

Marketplace sets per-token rates — renters choose priority tier when submitting jobs.

3. Run the Node.js Agent

Agent heartbeats keep your listing online and execute renter jobs.

Data center infrastructure

List Your Hardware

Provider account required to submit a listing.

Auto-detect from your node

Specs (GPU, CPU, RAM, TFLOPS) are read by the agent via nvidia-smi and Node.js — billing is per million input/output tokens in $NEXUS.

Step 1 — Start your node

Run this command on your GPU machine. The server assigns a unique node id when the agent connects.

Sign in to get your start command.

Earnings Calculator

$4,776/month

Provider Payouts (7 days)

Live settlement data

Live Node Metrics

WebSocket
Live Network

Network Statistics

Real-time metrics from the NexusCompute decentralized network.

—
Total GPUs
—
Available
—
Utilized
—
Network TFLOPS
—
Input/Output per 1M
—
Active Jobs
—
$NEXUS Price
—
FinOps Savings

Network Utilization (24h)

Network TFLOPS (24h)

Active Training Jobs (24h)

$NEXUS Token Price (24h)

Live Infrastructure Metrics

Real-time GPU telemetry streamed from provider nodes

Live
Developer API

API Documentation

Full REST + WebSocket API for AI agents to programmatically rent and run compute.

Quick Start

// Install renter SDK npm install @nexuscompute/renter-sdk // Submit job to online hardware — billed via settlement on completion const { NexusRenter } = require('@nexuscompute/renter-sdk'); const client = new NexusRenter({ apiKey: 'your_key' }); const job = await client.submitJob({ hardwareId: 'hw-a100-001' }, 'gemm', { a: [[1,2],[3,4]], b: [[5,6],[7,8]], backend: 'gpu' });
Open Swagger UI Download OpenAPI Spec

Key Endpoints

  • POST /api/v1/auth/login
  • GET /api/v1/hardware
  • GET /api/v1/hardware?online=true
  • POST /api/v1/jobs
  • POST /api/v1/ops/gemm
  • POST /api/v1/ops/backprop
  • POST /api/v1/ops/benchmark
  • GET /api/v1/network/stats
  • WS /ws (real-time metrics)
Transparent Pricing

Hardware-Class Token Rates

Rates vary by GPU tier and priority. All charges settle in $NEXUS when jobs complete.

$0.42
$NEXUS Price
$2.85M
24h Volume
40%
vs Traditional Cloud
15%
$NEXUS Holder Discount

How Billing Works

Submit jobs with a hardwareId — pricing follows that hardware's class. Priority tier sets your place within the rate range.

  • Input + output tokens billed per million at settlement
  • 2% platform fee — provider receives 98%
  • TTFT SLA: slow nodes get 50% payout slash; next job rerouted
  • Receipts verified peer-to-peer
$Nexus

$Nexus Wallet

Your $NEXUS balance for compute settlement.

$Nexus

Your account balance — compute jobs debit $NEXUS at settlement

—
$NEXUS Balance
—
$NEXUS Value (USD)
—
$NEXUS Rate

Wallet Activity

Sign in to view compute settlements.

Pay for Compute

Jobs settle in $NEXUS per million tokens from your $Nexus balance.

Priority Tiers

Higher tiers cost more $NEXUS but jump the queue.

Provider Earnings

Providers receive $NEXUS payouts when jobs settle.

Fund Your Wallet

Add $NEXUS to your balance before running compute jobs via the renter SDK.