LLMJob

Build your own AI infrastructure with spare GPUs and devices. Get OpenAI-compatible API access and monetize excess capacity when idle.

Features

• ⚡ Pool office workstations and spare hardware into a private AI cluster
• 🤖 Fully OpenAI-compatible API endpoints
• 📊 Real-time cluster monitoring dashboard
• 🔒 Keep sensitive data on your own infrastructure
• 💸 Turn idle GPU time into crypto today with the LLMJob Earn desktop app

This repository contains two packages:

• Server (repo root) — the Express API plus the static dashboard pages, backed by Postgres and deployed to Railway / GitHub Pages.
• LLMJob Earn (earn/) — an Electron desktop app that turns idle GPU time into crypto by mining Pearl (PRL) with its own CUDA miner: paste a payout address, hit Start, and earn — no command line. It's the easy on-ramp that gets GPUs onto the network today, ahead of LLM co-mining. See earn/README.md.

Add a node

Link a machine to your account with LLMJob Earn. Grab your pairing token from the Add Node page in the dashboard, then either:

• Desktop app — paste the token into the API → Connect tab and hit Connect.
• Headless CLI — run the command-line client on a server with no desktop:

llmjob-earn-cli connect --token <pairing-token>

Either way the machine generates its own key locally (only the public key is sent), joins your account, and pings so it shows online. See earn/README.md.

Run LLMJob Earn

Download the latest Windows installer from the releases page, install, paste a Pearl (prl1p…) payout address, and hit Start. To run it from source:

cd earn
npm install
npm start                          # launch the Electron app

See earn/README.md for the mining engine, the live balance, and building the installer.

Running the server

git clone https://github.com/super3/llmjob.git && cd llmjob
npm install                        # Install dependencies

npm start                          # Apply migrations, then start the server (default port 3001)
npm run dev                        # Same, with auto-reload (development)
npm test                           # Run test suite with coverage
npm run test:watch                 # Run tests in watch mode

The server requires a DATABASE_URL pointing at Postgres; migrations in server/migrations/ are applied automatically by npm start / npm run dev.