idle-node

idle-node

IDLE Protocol node — turn idle compute into USDC revenue on Solana.

Turn your machine into a paid node on the IDLE network. Run distributed compute tasks and AI inference, earn 85% of every job fee in USDC on Solana.

Install

npm install -g idle-node

Or run directly:

npx idle-node start --wallet YOUR_SOLANA_WALLET

Commands

idle-node start

Start the compute node and begin processing jobs (HTTP checks, scraping, monitoring, etc.).

idle-node start --wallet YOUR_SOLANA_WALLET [options]

Options:
  --types <types>         Comma-separated job types (default: all)
  --gpu                   Enable GPU-accelerated tasks
  --vram <gb>             VRAM to allocate (GPU mode)
  --poll-interval <ms>    Job poll interval (default: 5000)
  --max-concurrent <n>    Max concurrent jobs (default: 3)

idle-node inference

Start the inference worker for distributed AI inference using llama.cpp.

idle-node inference --wallet YOUR_SOLANA_WALLET [options]

Options:
  --llama-url <url>       llama-server endpoint (default: http://localhost:8080)
  --backend-url <url>     IDLE backend URL (default: production)
  --poll-interval <ms>    Job poll interval (default: 5000)
  --max-concurrent <n>    Max concurrent inference jobs (default: 2)

Quick start:

# 1. Start llama-server with a GGUF model
llama-server -m llama-3.2-3b.Q4_K_M.gguf --port 8080 -c 4096

# 2. Start the inference worker
idle-node inference --wallet YOUR_SOLANA_WALLET

See the Inference Setup Guide for full instructions.

idle-node status

Check node status and earnings.

idle-node status --wallet YOUR_SOLANA_WALLET

idle-node gpu-info

Detect GPU hardware and Ollama availability.

idle-node gpu-info

idle-node types

List all supported task types and pricing tiers.

idle-node types

Supported Task Types

Tier	Types	Price Range
Basic Monitoring	http_fetch, api_health, dns_lookup, ssl_check, response_time, json_validate	$0.001-0.002
Web Intelligence	html_scrape, element_extract, link_extract, price_extract, content_verify, change_detect	$0.003-0.005
GPU Compute	gpu_inference, gpu_embedding, gpu_info	$0.05-0.10
AI Inference	llama-3.2-3b, llama-3.2-1b (via llama.cpp)	$0.0005-0.001/256 tokens

Earnings

• 85% of every job fee goes directly to you
• Paid in USDC on Solana every ~5 minutes
• No minimum payout threshold
• Check earnings: idle-node status --wallet YOUR_WALLET

Requirements

• Node.js 18+
• Solana wallet address (for USDC payouts)
• For inference: llama.cpp with a GGUF model

Links

• Website
• Docs
• X / Twitter
• Telegram

License

MIT