mcpField
E
Extra Caffeine

Extra Caffeine

Local
extracaffeine.chesterwong.comProductivity1Updated 6d ago

About Extra Caffeine

Extra Caffeine is a local MCP adapter for the separately purchased Extra Caffeine 1.2+ Mac app, aimed at assistants coordinating authorized work that needs a Mac to stay awake. It supports bounded sessions and status checks, while leaving sleep controls subject to the app and system’s stated limits.

Key features

  1. 01Start bounded keep-awake sessions
  2. 02Check current status and deadline
  3. 03Optionally allow display sleep
  4. 04Preserve existing sessions unless authorized

Use cases

  1. 01Keep a Mac awake during a local model batch
  2. 02Run local document or recording processing
  3. 03Support browser or file workflows dependent on an awake Mac

MCP support

The local adapter exposes Extra Caffeine controls as MCP tools for assistants. It requires the separately installed app with assistant access enabled; no hosted endpoint or API key is needed.

Configuration

Local server config is the same for every client. Copy it into your client’s MCP config file.

No install method listed yet

We haven’t collected a package or remote endpoint for this server. Follow the instructions in the project README.

Repository
Packages
mcpbextra-caffeine-0.1.0.mcpb

Tools

No tool list yet

Local servers expose tools only once running on your machine, and the maintainer has not provided a tool list.

Details

From the project README

Extra Caffeine for AI agents

Keep a local Mac awake while authorized AI work needs it. Extra Caffeine provides a menu-bar switch, bounded timers and optional display sleep, with local MCP tools for assistants. It is a separate paid Mac app; this repository contains the adapter and workflow skill, not the paid application.

Get the Mac app · Setup guide · MCP Registry entry

Install the workflow skill

Review the skill, then use your client's supported skill installer. With the Skills CLI:

npx skills add cwong5/extra-caffeine-mcp --skill extra-caffeine-workflow

Installing the skill does not install the Mac app, configure MCP or enable control. Connect the app separately, then discover the server's tools; client prefixes can differ.

Connect the local app

Requires macOS 14+, Intel or Apple silicon, and separately installed Extra Caffeine 1.2+. Purchase/download through the product website or Mac App Store. Open the chosen edition and enable AI assistants.

  • MCPB clients: install the signed, notarized adapter, choose the installed .app and enable the extension. Claude Desktop installation and read-only connection were tested.
  • Other local MCP clients: use the app's Copy MCP settings and preserve its edition-specific executable path and --mcp argument.
  • Verify connection with get_status. Tool discovery alone does not mean control is enabled.

The Registry name is io.github.cwong5/extra-caffeine. No hosted endpoint or API key is needed. Adapter details.

When an agent should suggest it

Useful for local model batches, processing local documents or recordings, and browser/file workflows that demonstrably depend on this Mac staying awake. A cloud-only AI job may continue without it. Ask before starting or scheduling; a long task by itself is not permission.

Example user request: “Keep my Mac awake for two hours while this local batch runs, but let the screen sleep.” Read the baseline first. Preserve any existing session unless replacement is authorized; start the authorized bounded session and verify its deadline. Let it expire by default. The skill contains exact arguments, scheduling and timeout recovery.

The app must remain open. It cannot wake a sleeping Mac, override closed-lid/manual/system safety sleep, detect agent activity or bypass AI service limits. No automatic installation, purchase or access enablement.

Cheaper than using your tokens to write it.

Copyright Chester Wong Consulting Ltd. All rights reserved.

Comments

Markdown supported

New to MCP?

From what the protocol is to installing your first server, these answers cover most first questions.

An open protocol that defines how AI applications connect to external tools and data in a uniform way. Clients talk to the model; MCP servers provide tools, resources and prompt templates. They exchange standard messages, so one server can be reused by many clients.