Skip to content

About

Resources

Code of conduct

Security policy

Stars

10 stars

Watchers

6 watching

Forks

Repository files navigation

Ocean Network Compute Dashboard

A Next.js web application for monitoring and interacting with Ocean Network. It provides node operators, compute job consumers, and community members with tools to explore network health, run compute jobs, manage nodes, track rewards, and more.

Live app: dashboard.oncompute.ai


Features

  • Node Explorer – Browse and search all registered Ocean compute nodes with detailed stats, hardware specs, location, uptime, and benchmark history.
  • Run a Job – Step-by-step wizard to select a compute environment, configure resources (CPU, RAM, disk, GPU, duration) and pay for the job. This flow ends by opening the user's editor of choice, and running the job with the help of Ocean Orchestrator extension.
  • Inference – Launch model inference services from default models, custom Hugging Face models, or service templates, and manage running instances.
  • Run a Node – Guided setup flow for starting up a new Ocean compute node.
  • Global Stats – Network-wide analytics: total nodes, jobs, rewards, GPU popularity, and country distribution.
  • Account – Per-wallet hub: compute consumer and node owner activity, escrow funds (deposit/withdraw), subsidy, access lists, remote storage, and node auth tokens.
  • Subsidy Flow – Token subsidy mechanism backed by a Google Sheets ledger, Gemini AI verification, and an on-chain access list.
  • Web3 Wallet Support – Two independent paths: native wallets connect via EIP-6963 discovery + Ethers.js, and email magic link / passkey / Google OAuth go through Privy + Alchemy Account Kit to a smart contract account.
  • ENS Resolution – Display human-readable ENS names for wallet addresses.
  • Analytics – Dedicated analytics layer (summaries, rewards history, GPU stats) from the analytics backend.
  • Node Storage – Per-node persistent storage for compute consumers to upload files for using in compute jobs. Browse and create buckets on a node, upload/delete files, and control access via reusable on-chain access lists.

Tech Stack

Layer Technology
Framework Next.js 16 (Pages Router)
Language TypeScript
Styling SCSS / CSS Modules
Web3 Ethers.js v6, EIP-6963, Privy, Alchemy Account Kit
Ocean SDK @oceanprotocol/lib, @oceanprotocol/contracts
P2P libp2p
Data Fetching TanStack Query
Charts Recharts
Map Leaflet
Forms Formik + Yup
AI Google Generative AI (Gemini)
Package Manager Yarn 4 (Berry)

Getting Started

Prerequisites

  • Node.js version pinned in .nvmrc
  • Yarn 4 – enabled via Corepack: corepack enable

Install dependencies

yarn install

Configure environment variables

Copy .env.example and fill in the required values (see Environment Variables below):

cp .env.example .env.local

Run the development server

yarn dev

Open http://localhost:3000 in your browser.

Build for production

yarn build
yarn start

Scripts

Script Description
yarn dev Start Next.js development server
yarn build Build for production (NODE_ENV=production)
yarn start Start the production server
yarn lint Run ESLint
yarn lint:fix Run ESLint with auto-fix
yarn format Run Prettier on all TS/TSX files

Environment Variables

Create a .env.local file from .env.example (or configure your deployment environment) with the following variables. Unless marked optional, a variable is required.

App

Variable Description
NEXT_PUBLIC_APP_ENV "production" selects Base mainnet and production backends. Any other value (e.g. "development") selects Ethereum Sepolia and dev backends.
NEXT_PUBLIC_LEGACY_ESCROW_ADDRESS Optional. Address of the previous Escrow deployment on Base. When set, the escrow page shows a contract selector so users can view and withdraw funds left in it. Ignored on Sepolia.
NEXT_PUBLIC_OUTAGE_BANNER_ENABLED Optional. Set to "true" to show the outage banner on every page.
NEXT_PUBLIC_OUTAGE_BANNER_MESSAGE Optional. Banner text. Supports headings (#, ##, ### at line start), **bold**, *italic*, `code`, [link](https://...) and line breaks (\n). Raw HTML is shown as plain text.

Auth & Wallets

Variable Description
NEXT_PUBLIC_PRIVY_APP_ID Privy app ID used for login
NEXT_PUBLIC_ALCHEMY_API_KEY Alchemy API key for smart contract accounts
NEXT_PUBLIC_ALCHEMY_POLICY_ID Alchemy gas sponsorship policy ID
NEXT_PUBLIC_APP_URL Public app URL used as the Google login redirect (e.g. http://localhost:3000)
ALCHEMY_RPC_URL Production only. Base mainnet Alchemy RPC URL, used server-side and behind the /api/rpc proxy. In development the app uses a public Sepolia RPC.

Fiat On-Ramp

Variable Description
NEXT_PUBLIC_RAMP_API_KEY Ramp host API key for card top-ups
NEXT_PUBLIC_RAMP_GET_STATUS_INTERVAL Optional. Top-up status polling interval in ms. Defaults to 5000.
NEXT_PUBLIC_RAMP_GET_STATUS_MAX_TRIES Optional. Max top-up status polls. Defaults to 20.

Subsidy Flow

Variable Description
GEMINI_API_KEY Google Gemini API key used in the subsidy AI flow
SUBSIDY_GSHEETS_SERVICE_ACCOUNT_EMAIL Google service account email for Sheets access
SUBSIDY_GSHEETS_SERVICE_ACCOUNT_PRIVATE_KEY Private key for the Google service account, with line breaks escaped as \n
SUBSIDY_GSHEETS_SPREADSHEET_ID Google Sheets spreadsheet ID for the subsidy ledger
SUBSIDY_GSHEETS_SHEET_NAME Sheet/tab name within the spreadsheet
SUBSIDY_GMAIL_ADDRESS Gmail address used to send subsidy confirmation emails
SUBSIDY_GMAIL_OAUTH_CLIENT_ID Gmail OAuth client ID
SUBSIDY_GMAIL_OAUTH_CLIENT_SECRET Gmail OAuth client secret
SUBSIDY_GMAIL_OAUTH_REFRESH_TOKEN Gmail OAuth refresh token
SUBSIDY_ACCESS_LIST_ADDRESS Subsidy AccessList contract on the active chain that claimed wallets are added to
SUBSIDY_ACCESS_LIST_PRIVATE_KEY Private key of a wallet allowed to mint on that access list
NEXT_PUBLIC_SUBSIDY_PROVIDERS Comma-separated subsidy provider addresses for the active chain

Compute & Inference

Variable Description
NEXT_PUBLIC_GPU_LIST Comma-separated list of supported GPU models shown in the job and inference wizards
NEXT_PUBLIC_DEFAULT_MODEL_PEER_IDS Optional. Comma-separated node peer IDs whose service templates populate the default-model quick-start packages. When unset, the quick-start list is empty.
NEXT_PUBLIC_TEMPLATES_NODE_URL Optional. HTTP URL of the node that serves the service template catalogue. Defaults to the built-in default node.
NEXT_PUBLIC_VLLM_TAG Optional. Fallback Docker tag for vllm/vllm-openai. Defaults to pinned stable v0.28.0; models that require a newer architecture use their official model-specific image tag automatically.
NEXT_PUBLIC_LLAMACPP_IMAGE Optional. llama.cpp Docker image. Defaults to ghcr.io/ggml-org/llama.cpp.
NEXT_PUBLIC_LLAMACPP_TAG Optional. llama.cpp CPU image tag. Defaults to server.
NEXT_PUBLIC_LLAMACPP_TAG_CUDA Optional. llama.cpp CUDA image tag. Defaults to server-cuda.

The Custom Models configuration keeps the fallback per deployment but also exposes a per-model vLLM runtime selector. Its automatic setting selects qwen38-flash-next for Qwen/Qwen3.8-Flash-Next, glm53-flash for zai-org/GLM-5.3-Flash, and the configured fallback for other models.

Note: Variables prefixed with NEXT_PUBLIC_ are exposed to the browser. All others are server-side only and must never be committed or exposed publicly.


API Backends

The app connects to two backend environments, selected automatically via NEXT_PUBLIC_APP_ENV:

Environment (Incentive) API Analytics API
development https://api-dev.oncompute.ai https://analytics-dev.oncompute.ai
production https://api.oncompute.ai https://analytics.oncompute.ai

ENS resolution always uses: https://ens-proxy.oceanprotocol.com/api


Networks & Tokens

The active chain is determined by NEXT_PUBLIC_APP_ENV:

Environment Chain Chain ID
development Ethereum Sepolia 11155111
production Base Mainnet 8453

Token contracts:

Token Sepolia (dev) Base (prod)
USDC 0x1c7D4B196Cb0C7B01d743Fbc6116a902379C7238 0x833589fCD6eDb6E08f4c7C32D4f71b54bdA02913

Project Structure

src/
├── pages/          # Next.js pages (routes)
│   ├── index.tsx           # Home
│   ├── nodes/              # Node detail
│   ├── nodes/[nodeId]/storage/ # Node persistent storage (buckets & files)
│   ├── compute/            # Compute landing
│   ├── run-job/            # Compute job wizard
│   ├── inference/          # Inference models, templates, services & instances
│   ├── run-node/           # Node setup wizard
│   ├── stats.tsx           # Global network stats
│   ├── leaderboard.tsx     # Leaderboard
│   ├── account/            # Per-wallet account hub (activity, escrow, subsidy, access)
│   ├── subsidy/            # Subsidy flow
│   └── api/                # Next.js API routes (RPC proxy, Docker tags, subsidy)
├── components/     # Reusable UI components
├── context/        # React context providers
├── contexts/       # React context providers
├── hooks/          # Custom React hooks
├── data/           # Static data (e.g. inference quick-start packages)
├── services/       # Custom service functions
├── api-services/   # BFF API service functions
├── utils/          # Utility functions
├── types/          # TypeScript type definitions
├── constants/      # App-wide constants
├── styles/         # Global SCSS styles & design tokens
├── assets/         # Static assets (SVGs, images)
├── lib/            # Web3-specific wrappers & configs
└── config.ts       # App config, routes, and API route helpers

Links

About

Resources

Code of conduct

Security policy

Stars

10 stars

Watchers

6 watching

Forks

Releases

Packages

Used by

Contributors

Languages