Who is CueCrux designed for?

CueCrux is built for people who are answerable for what AI agents do.

Platform and infrastructure teams running agent fleets get the operational core: cost attribution that resolves from fleet to agent to verb to receipt with nothing estimated, and revocation that propagates as a signed pull with per-agent acknowledgements. When finance asks where the spend went or security asks whether access is really gone, the answer is a receipt, not a shrug.

Engineers debugging agents get receipts as a debugging tool. Instead of reconstructing an incident from scattered logs and model transcripts, you replay the signed trail of what the agent actually did, in order, with identity attached.

Compliance and governance teams facing the EU AI Act get controls that are the same machinery everything runs on: risk classes on plans and work items, human gates that cannot auto-approve past a timeout, hash-chained logging, and GDPR Article 17 forget with a dry-run. Engineering controls, not a legal opinion.

Individual developers get the free end of the same system: the Crux Daemon runs locally on CPU with no account, giving one person durable agent memory, budgeted retrieval and a verifiable record, at zero cost, forever.

The common thread is accountability. If nobody will ever ask you to produce a verifiable record of what your agents did, you may not need CueCrux. Most teams get asked.