A worker is the proving process. It dials the coordinator, registers a compute capacity, and proves the segments it's assigned. Workers are stateless across jobs — add, remove, or restart them at will. A worker host needs ~32 GB RAM and, optionally, a CUDA-capable GPU.
Reference
Usage
How to invoke zisk-worker, the default config path under the deploy script, override precedence, signal handling, and compute-unit sizing.
Read more →Reference
Configuration
worker.toml schema: identity, capacity, coordinator URL, connection tuning, logging.
Read more →Reference
CLI & environment
Every command-line flag (identity, keys, backend, performance, diagnostics), every ZISK_WORKER_* env var, and worked examples.
Read more →