PROVIDER://RUN NODE

SUPPLY COMPUTE.

Run the existing NexCompute worker on a dedicated Linux GPU host. The worker connects outbound to the coordinator, publishes detected GPU data, accepts eligible jobs and reports result receipts.

00 / REQUIREMENTS

Linux host with a supported NVIDIA GPU
Current NVIDIA drivers
Docker Engine
NVIDIA Container Toolkit for real GPU execution
A Solana wallet address used as the provider identity

01 / LOCAL MOCK MODE

TERMINAL 01

make api

TERMINAL 02

NEXCOMPUTE_MOCK_EXECUTION=1 make worker

Mock mode is for local development. It publishes a mock GPU and produces deterministic mock receipts. It must not be represented as real GPU processing.

02 / REAL GPU WORKER

export NEXCOMPUTE_API_URL=https://api.example.com
export NEXCOMPUTE_BOOTSTRAP_TOKEN=<provided-securely>
export NEXCOMPUTE_PROVIDER_WALLET=<solana-wallet>
export NEXCOMPUTE_PRICE_PER_HOUR_USDC=0.50
export NEXCOMPUTE_ALLOWED_REGISTRIES=ghcr.io,docker.io
export NEXCOMPUTE_MOCK_EXECUTION=0

cd apps/worker
go run .

The bootstrap token is operational authentication and must never be exposed in browser JavaScript, public configuration, screenshots or public documentation.

03 / EXECUTION POLICY

Production workers require digest-pinned images such as ghcr.io/organization/image@sha256:.... The current worker executes containers with no network, a read-only root filesystem, all Linux capabilities dropped, no new privileges, process limits, CPU and memory bounds, and a fixed maximum runtime.