Hey Product Hunt 👋 I'm César. OpenAdmiral started while I was building Applivery Intelligence at Applivery: an architecture where several agents interact with each other, with dependencies on other teams across the company. I ran a Claude Code agent for each piece, in parallel. And the bottleneck turned out to be me. Six terminals, six scrollbacks. The agent that needed a permission waited in silence behind a window I wasn't looking at. When one piece needed something from another, or from another team, I was the one copying context between terminals. So I built the tool I was missing: OpenAdmiral, a desktop bridge on top of Claude Code: 🧭 Every agent on one map: status, branch, context left and cost, grouped by repo 🔔 A notification the moment one needs you, and one click into its own terminal 🤝 Agents delegate to each other with a brief and report back as they go ⏭️ Up next: queue the next prompt to run when an agent is free, after another agent's work lands, or on a schedule 🌙 Underway: close the lid and they keep working. The Mac stays awake while agents have work, and the top right tells you when it's safe to close the lid 📚 Every result kept: reports, diagrams, UI mocks and HTML pages, versioned 💰 Costs per project and model, and how much of your plan is left A few things I care about: It's the real Claude Code. OpenAdmiral is a plugin, not a wrapper: your plugins, hooks, skills and MCP servers just work. Remove it and nothing changes. It's 100% local. Your code and sessions never leave your Mac, and there's no account. It's free and MIT licensed. It's an early alpha (macOS 13+), and if you live in the terminal: brew install --cask openadmiral/tap/openadmiral. I'd love your feedback: what would you need to run more agents at once? The site itself runs the real app in your browser as you scroll: https://openadmiral.com

OpenAdmiral

Mission control for your Claude Code agents

Mission control for your Claude Code agents