AgentOS Mini

When I started coordinating AI planners and executors, the models were not the main problem. The operating layer was.

Tasks lived in long chats. Status was hard to see. Approvals were easy to miss. If a process stopped halfway, recovery meant reconstructing context by hand.

I built AgentOS Mini to solve that problem with a small, local and inspectable workflow.

It includes:

• a structured task queue

• status updates and heartbeats

• approval gates for sensitive actions

• locks, timeouts and recovery procedures

• an append only financial ledger

• a hash chained audit log

• operating documentation and automated tests

AgentOS Mini is not an autonomous black box. It does not publish content, execute payments or create accounts on its own. It gives AI assisted work structure while keeping people responsible for critical decisions.

The first version is available for $29:

https://payhip.com/b/z6gFk?utm_source=indiehackers&utm_medium=post&utm_campaign=agentos_mini_launch

I would especially value feedback from founders and operators coordinating more than one AI workflow. What is the first coordination problem you would want a kit like this to solve?