🚀 Maximize your product's SEO. Submit to 240+ directories in 1-click with DirSubmit. Launch Now
WhoIsCooking logo

WhoIsCooking

See which AI agent is heating your Mac. Cool it down.

2026-10-26

Product Introduction

  1. Definition: WhoIsCooking is a lightweight macOS utility application (1.7 MB) designed for Apple Silicon Macs (M1 to M5) running macOS 14 or later. It functions as a real-time thermal and process monitor specifically tuned to identify AI-powered coding agents and development workloads.
  2. Core Value Proposition: It exists to solve the problem of thermal throttling and excessive fan noise on Apple Silicon MacBooks caused by intensive, often background, AI agent processes. Its primary value is providing granular visibility into which specific application or agent (like Claude Code, Cursor, or Codex CLI) is causing high CPU temperatures and offering a one-click "cooling" mechanism that intelligently re-prioritizes workloads without terminating them.

Main Features

  1. Real-Time Thermal & Process Monitoring: The app reads proprietary Apple Silicon chip temperature sensors (e.g., CPU, GPU) every 20 seconds, displaying the data in the menu bar or Dynamic Island notch. It correlates this thermal data with active processes, using heuristic analysis to group related child processes (like ffmpeg, cargo build, dev servers) under their parent AI agent session, presenting a simplified, single-line view per agent instead of dozens of cryptic process names.
  2. Intelligent Process Cooling (E-Core Shifting): The core technical feature is the "Cool down my Mac" function. When activated, it uses macOS's underlying nice and process priority APIs to dynamically shift the heavy computational threads of the identified hot processes from the high-performance (P-cores) to the high-efficiency (E-cores) of Apple's ARM architecture. This reduces heat and power draw, thereby lowering chip temperature (claimed ~14°C drop), while allowing the work to continue asynchronously.
  3. Agent-Specific Detection & Alerts: The app maintains a signature database to recognize specific AI coding tools (Claude Code, Codex CLI, Gemini CLI, Cursor, Aider, GitHub Copilot CLI, Ollama, LM Studio) and common development workloads (Xcode builds, Docker, browser tasks). It provides proactive notifications for "runaway" processes—defined as an agent sustaining high CPU for five minutes or leaving development servers running after its main session ends.

Problems Solved

  1. Pain Point: "Thermal Throttling" and "MacBook Overheating" during development, specifically caused by opaque, resource-intensive AI coding assistants that launch sub-processes users may not directly control. This leads to performance degradation, loud fan noise, and reduced hardware longevity.
  2. Target Audience: Software developers, data scientists, and engineers who actively use AI-powered coding assistants (Claude Code, Cursor, etc.) on Apple Silicon MacBook Pros or MacBook Airs. It's particularly relevant for users working on long-running tasks like compiles, renders, model inference, or local development servers.
  3. Use Cases: Essential when a MacBook fan suddenly spins up during a coding session; when a user needs to extend battery life during mobile work by reducing thermal load; when debugging performance issues to identify if an AI agent is the culprit; when wanting to run a background build/test while keeping the system responsive for other tasks like video calls.

Unique Advantages

  1. Strengths & Limitations (Pros & Cons):

    • Pros: Exceptional specificity in diagnosing AI-agent-induced heat. Non-destructive cooling preserves workflow state. Extremely low resource overhead (~0% CPU, 14MB RAM idle). One-time purchase ($5) vs. subscription. Apple notarized for security. Simplifies complex process trees.
    • Cons: Exclusively for Apple Silicon (M1+), no Intel Mac support. Cooling reduces performance speed for affected tasks. Requires macOS 14+. Is a monitoring/mitigation tool, not a root-cause fix for inefficient code.
  2. Key Alternatives & Differentiation:

    • Activity Monitor (Built-in): A generic system monitor. WhoIsCooking differentiates by agent-specific identification, thermal correlation, and one-click action (cooling). Activity Monitor only shows raw processes without grouping or thermal context and offers no mitigation.
    • iStat Menus, TG Pro: Advanced system monitors with temperature readouts. Differentiation: WhoIsCooking is hyper-specialized for AI/developer workflows, cheaper (one-time fee vs. ~$20), and offers the unique E-core shifting "cooling" action tailored to development workloads, not just monitoring.
    • Manual Terminal Commands (e.g., sudo kill, renice): Manual intervention is possible but risky and complex. WhoIsCooking provides a safe, GUI-driven, reversible method that avoids killing processes and requires no command-line expertise.

Frequently Asked Questions (FAQ)

  1. Does WhoIsCooking slow down my computer or AI agents? Yes, but intentionally and reversibly. The "cooling" function deliberately shifts work to slower efficiency cores to reduce heat. This can slow the specific agent's task completion. Full performance is restored instantly with the "Full speed" button or by quitting the app.

  2. Is WhoIsCooking safe to use? Will it crash my apps? Yes, it is safe. The app is signed with an Apple Developer ID and notarized by Apple, passing macOS security checks. It does not force-quit ("kill") applications unless you explicitly select "Quit." Its cooling function uses standard, low-level macOS APIs to adjust process priority, which is non-destructive and should not cause crashes.

  3. Can WhoIsCooking cool down an Intel-based Mac? No. The application relies on reading temperature sensors unique to Apple Silicon chips (M1, M2, M3, M4, M5). On Intel Macs, it cannot display temperature data, which is core to its functionality, though it could theoretically still show process groupings.

  4. How does WhoIsCooking detect which AI agent is running? It uses a combination of process name matching, parent-child process tree analysis, and heuristic behavioral patterns to identify known AI agent executables (like claude-code, cursor-agent) and intelligently group their spawned sub-processes (build tools, servers) into a single, understandable line item.

  5. What happens if I quit the WhoIsCooking app? All process management performed by the app is reverted. Any workload shifted to efficiency cores is returned to normal priority, restoring full system performance. The app leaves no persistent changes to your system's configuration.

Submit to 240+ Directories with 1-Click

Maximize your product's SEO and drive massive traffic by automatically submitting it to over 240 curated startup directories using DirSubmit.

Related Products

Subscribe to Our Newsletter

Get weekly curated tool recommendations and stay updated with the latest product news