WakeUp Samurai

WakeUp Samurai started from a practical annoyance: an agent is coding, I step away from the computer, and macOS goes to sleep at the wrong moment. The app lives in the menu bar, detects active agent processes, and keeps the Mac awake only when that is actually needed.

Classic stay-awake apps rely on manual toggles. With AI coding agents, that is easy to miss: the Mac can sleep during a task, or stay awake long after the task is done.

If the app understands when an agent is really working, it can behave better than a manual switch: keep the machine alive during the job and hand control back to macOS when the agent stops.

• Native macOS menu bar application.
• Process detection for coding agents including Codex, Claude Code, Cursor, Gemini, Copilot, Kimi, and JetBrains AI.
• Automatic power assertion only while agents are actively working.
• Launch-at-login settings and manual control from the menu.
• Public open-source repository plus a separate landing page with a download CTA pointing to the latest release.

AI is not a chatbot inside the interface here. The product is built around real AI-agent workflows: it watches the state of agent tools and reacts to them instead of asking the developer to manage sleep settings manually.

Swift, SwiftUI, AppKit, process monitoring, IOKit power assertions, Launch Services, and a separate Next.js landing page deployed on Vercel.

• Start with a small native tool that solves one annoying workflow problem.
• Use the menu bar instead of a full app because the product should stay out of the way.
• Make it open-source from the beginning so developers can trust the tool that keeps their Mac awake.
• No telemetry and no App Store dependency for the first release path.

A public project with a GitHub release, a Vercel landing page, and broad detection coverage for agent tooling.

Tighten distribution, collect feedback from developers using coding agents, and add more integrations only when real usage justifies them.