Nodus logo
Powered by ICPX Protocol

NODUS

AI  Compute  Coordination  Layer

High-performance AI compute, made accessible. Nodus is the coordination layer that connects developers, agents and applications to decentralized GPU infrastructure — through one simple, scalable interface.

Launch App Read Docs
Built on ICPX PROTOCOL
SCROLL
0
GPUs Online
0
Jobs Running
0ms
Median Latency
0%
Network Uptime
The Problem

AI builders struggle with GPU availability, runaway cost management, deep infrastructure complexity, and brittle deployment workflows.

The Solution

One unified dashboard for compute orchestration, workload execution, model testing and agent runtime — abstracting decentralized GPUs into a single clean interface.

Core Products

Everything to ship AI compute

Compute Hub

Discover, reserve and orchestrate GPUs across the network from a single command surface.

Agent Runtime

Deploy and run autonomous AI agents with persistent, schedulable execution environments.

Model Playground

Test, benchmark and compare models interactively before committing them to production.

Usage Analytics

Transparent, real-time accounting of compute spend, throughput and resource utilization.

API Access Layer

A single programmable surface for compute, agents and models — integrate Nodus into any app, pipeline or autonomous system with a few lines of code.

Architecture

From app to silicon, coordinated

01
Frontend Layer
Apps, agents & developers
02
Nodus Coordination
Scheduling & orchestration
03
ICPX Infrastructure
Compute settlement on Solana
04
GPU Providers
H100 · A100 · RTX 4090
Developer API

Ship a job in
five lines

Schedule workloads on decentralized H100s without touching infrastructure. The Nodus SDK handles discovery, provisioning, billing and teardown.

Real GPU access, on demand
Pay-per-second compute scheduling
Native agent & model endpoints
run_job.ts
import { Nodus } from "@nodus/sdk";

const nodus = new Nodus({ apiKey: process.env.NODUS_KEY });

// schedule on a decentralized H100
const job = await nodus.compute.run({
  gpu:     "H100",
  image:   "vllm/vllm:latest",
  command: "python serve.py --model llama-3-70b",
});

console.log(job.endpoint); // https://…nodus.run
Compute Hub

Your control room for compute

app.nodus.run/compute
NODUS
Compute Hub
Agent Runtime
Playground
Analytics
API Keys
Balance
142.6 SOL
GPU Availability
Live across the ICPX network
LIVE
H1003.2k free
2.40 SOL/hr
A1005.1k free
1.15 SOL/hr
RTX 4090820 free
0.34 SOL/hr
UTILIZATION · 24H
ACTIVE JOBS
llama-3-70b
running
sdxl-batch
running
agent-swarm
queued
whisper-v3
running
Security

Secure by architecture

Secure Execution

Isolated, verifiable workload execution across every node in the network.

Transparent Accounting

Every unit of resource is metered and settled on-chain — no hidden costs.

Infrastructure Resilience

Decentralized by design — no single point of failure, automatic failover.

The coordination layer
for AI compute

Connect your applications to decentralized GPU power. Start building on Nodus today.

Launch App ↗ Read Docs