AgentMicro is a local-first macOS menu-bar companion for supervising parallel Codex Desktop and CLI tasks. It surfaces task state, project, elapsed time, and results at a glance, then opens the matching Codex Desktop task with one click. It observes local metadata only and never uploads prompts, responses, source code, or task history.
Loading comments…
Project Info
Product Keywords
AgentMicro is a local-first macOS menu-bar companion built for developers who run multiple Codex Desktop and CLI sessions at once. It gives you a live overview of every parallel task—showing state, project, elapsed time, and results—without forcing you to switch windows or reopen Codex. One click on any task in the menu opens the matching Codex Desktop conversation directly. AgentMicro reads only local session metadata, so your prompts, responses, source code, and task history never leave your machine.
AgentMicro surfaces thinking, unread results, input requests, errors, and idle tasks directly in the menu bar. Working tasks stay pinned first, so you always know which session needs your attention right now.
Each row in the queue includes the project name, current-turn duration, and a fast-mode marker. This extra context helps you decide where to return without opening each conversation to check.
Clicking a task in the AgentMicro menu opens the matching Codex Desktop conversation instantly. No copy-pasting IDs, no searching through window lists—just a direct jump from overview to action.
Base mode requires no Accessibility permission and no account. Optional enhanced detection is explicitly enabled, reads visible local labels only, and never clicks, types, or approves anything on your behalf. AgentMicro is independent, community-maintained, and not affiliated with OpenAI or CodexBar.
AgentMicro gives you a full picture of every Codex task without ever seeing your code.
That privacy-first stance is the core differentiator. Most task managers or AI companions require cloud sync or broad system access; AgentMicro deliberately works with local metadata only. It's free and open source, runs on macOS 14+, and supports both Codex Desktop and CLI—so you get a genuinely useful supervision layer without compromising on data control or adding another subscription.
You regularly run multiple Codex sessions in parallel and lose track of which one is waiting for input, which has results ready, and which is still thinking. If you value a local-first tool that respects your privacy by default, and you want a lightweight menu-bar companion rather than a full project-management suite, AgentMicro fits naturally into your workflow. It's also worth a look if you prefer open-source, community-maintained tools over commercial alternatives with opaque data handling.
Other tools you might consider
Emdash is an open-source desktop app for running multiple coding agents in parallel; one place to monitor sessions, review diffs, and turn issues into PRs.
Run background coding agents from anywhere. Spawn Claude Code or Codex in a VM with code and tooling ready to go. Hand off tasks from Slack, Linear, or GitHub Replicas runs Claude Code, Codex, or any coding agent in the cloud. Agents run in isolated VMs with real dev environments, and you can bring your own subscriptions and API keys. Trigger tasks from Slack, Linear, or GitHub and come back to a PR ready to review.
A macOS workbench where Claude, Codex, and Hermes run on your specs, research, and Figma files.
Three things kept hitting me using Claude Code. Tab switching every 30 seconds just to check if it's still running. Claude silently blocking for 12 minutes while I was in another app. Coming back to find it finished or stuck 15 minutes ago. So I built CodeBreak. A pixel-art character walks your screen while CC runs. It celebrates when done, panics when it needs you, sulks on errors. $7 one-time. No sub. All future updates are free. Its for CC now, eventually will be Universal for all AI tools.
Maker
kettle_dev
Loading comments…