Skip to main content

Native worker overview

Native workers run Qwen2.5 14B (abliterated/uncensored) on your machine using node-llama-cpp. They serve Max tier requests and earn 3-5x more than browser workers.

Why go native?​

  • Higher earnings — Max tier jobs pay significantly more than Free or Pro
  • Better model — 14B parameter model produces higher quality responses
  • Real GPU utilization — uses CUDA, Metal, or Vulkan for full hardware acceleration
  • Runs headless — no browser tab needed, runs as a background process or service

Requirements​

  • Node.js 18+ (22+ recommended)
  • GPU with 10GB+ VRAM:
    • NVIDIA: GTX 1080 Ti, RTX 3060 12GB, RTX 3080, RTX 4070+, etc.
    • Apple Silicon: M1 (16GB RAM), M2, M3, M4
    • AMD: RX 6800+, RX 7800+ (via Vulkan)
  • ~10GB disk space for the model (downloaded on first run)
  • Stable internet connection

Quick start​

npx @c0mpute/worker --token <your-token>

That's it. One command. node-llama-cpp handles model download, GPU detection, and compilation automatically.

Get your token​

  1. Go to c0mpute.ai/worker
  2. Login with Privy
  3. Scroll to Native Worker section
  4. Click Get Worker Token
  5. Copy and save the token — it's shown only once

See Worker tokens for more details.

Platform guides​