Skip to content
Stork.AIstork.ai
AI Tool

AgentPeek Review

AgentPeek is a macOS application designed for real-time monitoring and interaction with AI coding agents like Claude Code and Codex, directly from the Mac notch and menu bar.

shipped May 10, 2026updated May 27, 2026aifreemium
aicoderesearch
AgentPeek — product screenshot

Why it matters

1AgentPeek operates on macOS 14 or later, specifically on Apple silicon Macs.
2The application provides real-time token usage tracking for Claude Code and Codex, with 5-hour and 7-day statistics.
3As of July 1, 2026, AgentPeek is in version v0.2.7, with frequent updates from solo developer Bren Huber.
4It has garnered 165 upvotes and 313 followers on Product Hunt.

Stork’s verdict on AgentPeek

AgentPeek offers glanceable real-time monitoring from your Mac notch, but it's limited to macOS and specific agents like Claude Code.

AgentPeek reviewed by Stork AI · stork.ai/en/agentpeek

About AgentPeek

Business Model
Subscription SaaS
Headquarters
New York, USA
Platforms
macOS
Target Audience
Developers and Mac users

Leadership

Bren Hubr

overview

What is AgentPeek?

AgentPeek is an AI observability tool developed by Bren Huber that enables macOS developers to monitor and interact with AI coding agents such as Claude Code and Codex. It provides real-time insights into AI agent sessions, permissions, token usage, and local development servers directly from the Mac notch and menu bar, without requiring users to switch applications. The application is designed for Apple silicon Macs running macOS 14 or later, offering a local-first approach where all session data remains on the user's device, ensuring privacy. AgentPeek aims to streamline the development workflow by providing glanceable status updates and interaction points for AI coding assistants, helping developers understand agent orchestration, execution traces, and cost attribution.

features

Key Features of AgentPeek

AgentPeek integrates directly into the macOS environment to offer a suite of monitoring and interaction capabilities for AI coding agents. Its features are designed to provide developers with immediate, actionable insights into their AI-assisted coding processes.

  • Live Session Monitoring: Displays active and recent AI agent sessions, including state (executing, thinking, waiting, idle), most recent activity (file edited, command running, plan adopted), and project name.
  • Permission and Question Handling: Allows direct responses to agent permission requests and freeform questions (e.g., Approve, Deny, Always Allow) from the notch panel.
  • Token Usage Tracking: Provides real-time visibility into 5-hour and 7-day token usage reported by Claude Code and Codex, including daily statistics.
  • Local Development Server Management: Automatically detects and lists local development servers running on ports 3000-9999, identifying common frameworks and runtimes, with options to open, copy, or stop URLs.
  • Quick Routes: Offers shortcuts to access a session's skills, plugins, MCP servers, configuration, or logs directly from the notch.
  • Debugging and Observability: Helps understand agent orchestration, execution traces (tool calls, inputs/outputs, retries, failures), prompts and results, and cost attribution.
  • Session Replay: Tracks files touched by agents and provides post-session debugging capabilities.
  • Fast Actions: Allows saving and running shell commands directly from the notch (introduced in v0.2.5).
  • Privacy-Focused: All session data is stored locally on the user's device, with no accounts, analytics, or telemetry.

use cases

Who Should Use AgentPeek?

AgentPeek is primarily designed for developers and Mac users who frequently utilize AI coding agents like Claude Code and Codex. Its functionality caters to individuals seeking enhanced observability and streamlined interaction with their AI assistants within the macOS environment.

  • Developers using Claude Code and Codex who require real-time monitoring of their AI agents' activities and status.
  • Mac users seeking efficient session management and quick interaction with AI agent prompts without switching applications.
  • Users needing comprehensive insights into AI model sessions, including execution traces, prompts, and results for debugging.
  • Developers tracking token usage and estimated costs for AI models to manage resource consumption.
  • Individuals monitoring local development servers related to AI models, requiring quick access and control from the menu bar.

how to use

How to Use AgentPeek

To begin using AgentPeek, users install the macOS application on an Apple silicon Mac running macOS 14 or later. Once installed, it integrates into the Mac notch and menu bar, automatically detecting compatible AI coding agents.

  • 1Install AgentPeek from the official website (agentpeek.app) on a compatible macOS device.
  • 2Launch the application; it will appear in the Mac notch and menu bar.
  • 3Initiate a session with a supported AI coding agent, such as Claude Code or Codex.
  • 4Monitor live session status, agent activities, and token usage directly from the notch panel.
  • 5Respond to AI agent permission requests or questions via the interactive prompts in the notch.
  • 6Access local development server information and quick routes to session details from the menu bar interface.

pricing

AgentPeek Pricing & Plans

AgentPeek operates on a freemium business model, offering a free tier with core functionalities and a subscription tier for advanced features. Specific pricing figures for the subscription tier are not publicly detailed beyond its existence.

  • Free Tier: Provides access to fundamental monitoring capabilities for Claude Code and Codex.
  • Subscription Tier: Offers additional features and enhancements, available through a paid subscription.

Pros

  • +Provides real-time, glanceable monitoring of AI coding agents (Claude Code, Codex) directly from the Mac notch and menu bar.
  • +Enhances developer workflow by allowing direct interaction with agent permissions and questions without switching applications.
  • +Offers detailed token usage tracking (5-hour, 7-day, daily statistics) for cost management and resource awareness.
  • +Includes local development server management, automatically detecting and listing servers for quick access and control.
  • +Ensures user privacy with a local-first approach, storing all session data on the device with no telemetry or accounts.
  • +Actively developed by Bren Huber, with frequent updates (e.g., v0.2.7 as of July 1, 2026) indicating ongoing improvements.

Cons

  • −Currently limited to macOS platforms, excluding Windows, Linux, or web-based users.
  • −Primarily supports Claude Code and Codex, with less explicit support for a broader range of AI coding tools compared to some alternatives.
  • −Specific pricing details for the subscription tier are not transparently published, making cost evaluation challenging.
  • −As a relatively new tool (initial release v0.1.0 in April 2026), its long-term stability and feature set are still evolving.
  • −Relies on the Mac notch design, which may not be universally preferred or available on all Mac models.

Similar Tools

AgentPeek vs Competitors

AgentPeek positions itself as a native macOS application providing real-time observability for AI coding agents, with a distinct focus on integration into the Mac notch and menu bar. This offers 'glanceable' status updates without workflow interruption, differentiating it from broader AI automation platforms or retrospective analysis tools.

1
SessionWatcher↗

It is a native macOS menu bar app specifically designed for real-time monitoring of Claude Code and Codex usage, including token counts and estimated costs.

Like AgentPeek, SessionWatcher offers real-time token usage and cost tracking for Claude Code and Codex from the macOS menu bar, directly competing on UI and core functionality. AgentPeek specifically highlights its integration with the 'Mac notch', while SessionWatcher focuses on the general 'menu bar'.

2
AIUsageOn StorkCompare

It is an open-source, local-first usage tracker that monitors tokens, cost, and sessions across over 20 AI coding tools, including Claude Code and Codex, with a local dashboard and desktop widget.

AIUsage shares AgentPeek's local-first and privacy-focused approach to monitoring AI coding assistant usage, providing a desktop widget for quick glances similar to AgentPeek's Mac notch integration. It supports a broader range of AI coding tools than AgentPeek's stated Claude Code and Codex focus.

3
codachiOn StorkCompare

It provides a context window monitor for Claude Code, displaying burn rate and time remaining, with an interactive ASCII pet that reacts to the user's workflow.

Similar to AgentPeek, codachi offers real-time, context-specific monitoring for Claude Code usage, focusing on token limits and remaining capacity. Its unique ASCII pet UI element provides a distinct visual feedback mechanism compared to AgentPeek's Mac notch integration.

4
AI Usage Measurement Framework↗

This open-source framework analyzes local Git commit history to detect and quantify AI-assisted development across multiple tools, including Claude, with a privacy-friendly approach.

While AgentPeek offers real-time session monitoring from the Mac notch, the AI Usage Measurement Framework provides a retrospective, local-first analysis of AI code adoption and usage trends by examining Git history. It focuses on broader team-level insights and ROI rather than individual real-time developer feedback.

Visit AgentPeek↗
Connect
𝕏
X / Twitter@brenhubr

AI Reputation Report

Is AgentPeek yours?

ChatGPT, Perplexity, Gemini, Claude & Grok answer buyer questions about AgentPeek every day. See whether they name AgentPeek — or send buyers to a rival.

See what AI saysfree preview