Turn idle GPUs into income

Universities, labs and studios: list spare NVIDIA or Ascend capacity on n8bee and earn per GPU-hour while it would otherwise sit idle.

Outbound only

The runner connects out over HTTPS. No open ports, no inbound SSH, works behind NAT.

Scoped access

Jobs get short-lived URLs for exactly the data they need. Revoke your pool token any time.

You set the price

Choose a rate per GPU-hour, the workloads you accept and the hours you're available.

How providing works

1
Create a pool
Register your lab box or cluster and receive a one-time enrollment token.
2
Start the runner
Run the n8bee runner container with GPU access. It detects your device, VRAM and drivers.
3
Pull jobs
The runner claims matching jobs, streams progress back and uploads results when done.
4
Get paid
Every completed GPU-hour is metered to your pool and paid out at your listed rate.

docker run -d --gpus all --restart unless-stopped \
  -e N8BEE_POOL_TOKEN=pool_... \
  n8bee/runner:cuda

Tell us what hardware you have. Early providers are onboarded first.