What is CueCrux and what does it do?

Your agents did 4,183 things today. CueCrux exists so you can prove every one of them.

CueCrux is the governance control plane for fleets of AI agents. It does three things that dashboards and log aggregators do not. First, cost attribution: every unit of spend resolves from fleet to agent to verb to the receipt that recorded it, with nothing estimated. Second, receipts as debugging: when an agent misbehaves, you replay the signed trail of what it actually did rather than reconstructing events from scattered logs. Third, revocation that propagates: pulling access is a signed operation with per-agent acknowledgements, and when zero holders remain you get an attestation saying so.

The foundation is the Crux Daemon: one free binary that runs locally, CPU-only, with no account. It gives your agents durable memory and retrieval, and it signs a CROWN receipt for every state mutation. Receipts are hash-chained on an append-only spine, so the record cannot be quietly edited, and verification works offline.

Identity runs on passports. Every connection binds to one, every tool call carries it, and verified receipts accrue as reputation across five trust tiers. Policy runs on RCX: short-lived capability tokens scoped per request, with refusals that fail closed and leave a signed RefusalReceipt.

The result is a fleet you can answer for. Any receipt resolves back to who acted, at what tier, under which grants. That is the whole product: receipts over vibes.