For the complete documentation index, see llms.txt. This page is also available as Markdown.

Welcome to Ocean Network

Ocean Network solves the "Coordination Problem" of Decentralized Compute.

Currently, the world has massive amounts of idle computing power (GPUs/CPUs) and data scientists who need it, but no easy way to connect them.

Ocean Network bridges this gap by providing:

  • For Node Operators (Supply): A way to monetize their hardware. The Dashboard and Ocean Node software let them "rent out" their machines, track their earnings, and compete on leaderboards based on performance benchmarks.

  • For Developers/Data Scientists (Demand): A "Catalog" of compute resources. They can easily filter for the exact specs they need (e.g., "I need an Nvidia H100 with 64GB RAM"), pay for it securely, and run their jobs (via tools like the VSCode extension) without needing a centralized cloud provider.

In short: It turns the entire global network of individual computers into one giant, accessible, and monetizable supercomputer for AI and data processing.

Inference: the network's flagship capability

Compute jobs run once and exit. Inference keeps a workload up: a model or an AI application running on a node's GPU, reachable over HTTP for as long as you have paid for it.

From the dashboard you can serve any model on the Hugging Face Hub behind an OpenAI-compatible endpoint, launch a curated package that is already sized and tuned for the hardware, or start a ready-made application such as ComfyUI, Open WebUI or JupyterLab. Placement, payment and lifecycle are handled by the network; you pay a node operator directly for the hardware and the minutes you use.

Key Concepts

  • Inference: Long-running models and AI apps on network GPUs, billed for the runtime you book. Curated packages, any Hugging Face model, ready-made services, or templates that ship their models with them.

  • Ocean Nodes: The backbone of the network. Nodes process compute jobs, index metadata, and store data. Anyone can run a node and earn rewards.

  • Ocean CLI: A command-line tool for developers and data scientists to interact with the network (publish, consume, run compute jobs).

  • Dashboard: A visual interface to monitor network health, track your node's performance, and run compute jobs.

  • Base Network: Ocean is deployed on Base (Ethereum L2) for low-fee, high-speed transactions.

Getting Started

Choose your path to interact with the Ocean Network::

Image
  • Run a Node: Support the network, earn rewards, and provide compute. Go to Ocean Node Guide

  • Explore all Network fetaures: Publish datasets, algorithms, and orchestrate complex compute flows from your terminal. Go to Ocean CLI Guide

  • Develop Algorithms & Run Compute Job: Write and test algorithms directly in your IDE. Go to VS Code Extension or other IDE based on VS Code and orchestrate complex compute flows from your terminal. Go to Ocean CLI Guide

Run a Model

Serve a model or launch an AI app on a node GPU

Run a Job

Orchestrate complex compute flows

Run a Node

Support the network and earn rewards. Go to Ocean Node Guide

Last updated