Operator Onboarding
How a machine joins the B3IQ fleet — invite-based onboarding, node registry state, owner binding, and route/runtime readiness.
B3IQ controls what enters the fleet: joining today is invite/manual onboarding, not a self-serve signup. A new machine enrolls with an onboarding code the B3IQ team issues; a machine that was pre-onboarded on your behalf is linked to your account by redeeming a claim code from your dashboard. Once a machine is onboarded and claimed, there is no separate approval step — it can serve routed work as soon as it is online and route-ready, and it earns on any B3IQ-directed work once it passes route, runtime, and model readiness. There is no token to lock and no minimum balance to post.
Node Registration
A node that is eligible to earn needs:
The node ID is registered in the node registry and bound to an owner.
Registry state marks the node active.
Registration is necessary but not enough; route, runtime, model, and benchmark gates still need to pass before the node is route-ready.
B3IQ keeps node and model registry state — node ID, owner binding, and active status — so readiness and payouts attach to an owner you control. Public docs and explorer views keep this state explicit.
Model Hosting
B3IQ separates model publisher commitments from node hosting capability:
- Model publisher commitments can include artifact hash, metadata URI, publisher status, and model-level policy boundaries.
- Node model-hosting capability can include node/model/runtime/benchmark/route commitments.
- Raw inventory, endpoints, filesystem paths, prompts, and outputs stay out of public commitments.
Capability Commitments
The target b3iq-capability-v1 commitment is public-safe. It can hash:
- Node ID.
- Model class.
- Runtime class or runtime hash.
- Benchmark hash.
- Route class.
- Privacy class.
- Metadata URI.
It must not include raw local endpoints, runtime IDs, paths, private routes, prompts, outputs, or secrets.
Enforcement Boundary
Withholding earnings is a last resort. Refund, release, withhold, suspension, and reputation paths should handle normal operational failures. Any enforcement action needs clear evidence standards, appeals, and public-safe reason strings.
See Reputation for dispute policy.