You're running more coding agents than ever, but you can't keep up with them. That's where AgentPeek comes in. It pulls every session up into your Mac notch, live. Glance up, approve a prompt, watch token usage and manage the entire flow without pausing your YouTube video. All local, all yours.
Loading comments…
Project Info
Product Keywords
AgentPeek is a macOS app that lives in your Mac notch and keeps your AI coding agents visible at all times. Instead of juggling multiple terminal windows or missing when an agent needs your input, AgentPeek surfaces live sessions, permission prompts, token usage, and local dev servers in one glanceable spot. It runs entirely on your own Mac, with no accounts, analytics, or telemetry—just a clean view of what your agents are doing.
Every Claude Code and Codex session appears side by side in your Mac notch, showing which ones are working, waiting for your input, or finished. You can see the entire state of your agents at a glance, without leaving your editor or pausing your video.
AgentPeek monitors your 5-hour and 7-day token limits for both Claude and Codex, each with its own reset countdown. You'll see a rate limit coming long before it interrupts your workflow, so you can plan your usage accordingly.
When an agent needs approval—plans to confirm, permissions to grant, or questions to answer—the request appears right in the notch. You can respond with a single keystroke: ⌘A to allow, ⌘N to deny, or ⌥A to always allow in Claude Code.
Local dev servers, session config, logs, MCP servers, and tool call history are all one click away from the notch. Every tool call is summarized in plain language with timestamps, so you can see exactly what your agents touched.
AgentPeek actually knows what your agents are doing, instead of just showing a terminal.
Most agent monitoring tools only display raw terminal output, but AgentPeek follows your Claude Code and Codex sessions intelligently. It surfaces permissions, token usage, diffs, and local servers sorted by project, all without requiring you to change how you work. The notch keeps whatever needs your attention front and center, while the menu bar offers the same view if you prefer it there.
You run multiple coding agents simultaneously and want to stay on top of their activity without constant context switching. AgentPeek is especially useful if you value privacy (everything stays local, no telemetry), need to track token limits proactively, or want to approve agent prompts with keyboard shortcuts while keeping your focus elsewhere. The free 2-day trial gives you a real run before a one-time $15 purchase unlocks it permanently on your Apple silicon Mac.
Other tools you might consider
Conan is a native macOS app that wraps Claude Code in a live HUD — every prompt, tool call, skill, and token, surfaced as it happens.
Git for your AI agent’s actions. Undo, trace, and control every step. re_gent shows what your coding agent changed, which prompt caused it, and lets you roll back agent work across files and sessions.
Evaluate AI agents before they fail. Create test suites, run evaluations, and pinpoint issues before they reach production. AgentX provides full observability and traceability for your AI agents. AI analysis not only identifies problems but also suggests fixes-like an AI doctor for your agents. Simulate run your agents across multiple LLM providers to compare performance, cost, and latency, helping you make better decisions about which LLM to go. Run eval before deploy. Like CI/CD for AI agents.
Your MacBook's notch just sits there doing nothing. Now it takes notes. Hover over it, a panel drops down, you jot the thing down and carry on. And yes, there's a mildly passive-aggressive mascot.
Maker
sleepyfox
Loading comments…