What it does
OpenAdmiral has two parts. A plugin runs inside every Claude Code session and publishes its state. The macOS app reads that state and shows every session from every project in one window. Each session shows its live status, how much context it has left and what it costs. When an agent stops for a permission or a decision, the app sends a native macOS notification. Clicking it takes you to that agent, where you can answer. Agents can hand work to other agents with a brief and report back with a status. You can queue prompts and slash commands to run later. The app also keeps each conversation readable and collects the reports, plans, diagrams and pages that agents produce. OpenAdmiral calls itself an early alpha.
Capabilities
- Queued prompts and slash commands run when an agent is free, once another agent's work lands, or on a recurring schedule such as weekdays at 08:00.
- The Underway mode keeps a Mac awake while agents have work. It stops by itself below a battery floor of 20% by default, if the Mac gets too hot, or after a time limit of 8 hours by default.
- Each session's whole conversation is searchable, including every prompt, tool call and reply.
- Reports, UI mocks, diagrams and live HTML pages are collected in one gallery with every version kept.
- Cost views show spend per project and per model, a month-end projection, and 5-hour and weekly plan limits with reset times.
When to use it
- You run several Claude Code sessions at once and want to see which are working, finished or stuck.
- An agent may be waiting on a permission or decision while you are looking at another window.
- You want agents to keep working on queued or scheduled prompts while you are away from your desk.
