Agents on their infrastructure. Control in their hands.
I built Fleetnook to explore the handoff between an agent developer and a customer running it: approved Docker deployments, runtime health and signed usage reporting in one control plane.
- Constraint
- An agent developer needs enough operational signal to help, while the customer stays in control of their environment.
- Decision
- Use an outbound customer-side runner, customer-approved deployments and revocable, expiring diagnostic access.
- Evidence
- Independent demo · invitation-only hosted beta. Built a hosted control plane and customer-side runner with deployment approval, signed telemetry and recovery after connection loss.
Read more about the work
Built the hosted control plane and customer-side runner, with deployment approvals, signed telemetry, recovery after connection loss and protection against duplicate usage reports.
The runner connects outbound from the customer environment. Fixed diagnostics require an expiring grant that the customer can revoke; the customer chooses their own Temporal service.
An independent demo with an invitation-only hosted beta. Interactive support, managed Temporal hosting and payment collection remain future work.
Independent product / Fleetnook
A control plane with a clear boundary.
I built the console and customer-side runner together: approve the deployment, see runtime health and inspect reported usage.
View full size ↗Customer approval before the runtime starts.
Health and signed usage reports from the runner.
Expiring, revocable diagnostic access.











