Prototype · Preview: every server in the official registry, synced every 30 minutes. Reviews and accounts come in a later step.
PopNx MCPMCP servers for people who use AI · explained in plain words

mozak

by pitfa19version 0.4.5i★ –iRuns on your computer
Registry nameio.github.pitfa19/mozak

Install

Overview

Description from the maker (original English)

One owner-approved plan per project, shared by coding agents as plain files, with an Obsidian view.

PopNx MCP’s three pillars

These three are never merged into one score: opinions, measurements and popularity are different things.

What you can do with it

We have not written a plain-language summary for this one yet. Below are the facts we read from the registry.

  • Connection: Runs on your computer
  • Transport: stdio
  • Category: Docs, sheets & office work

Install

Pick the AI app you use, then copy the settings.

Before you install

  • Basic checks: 1 of 1 passed
  • Runs on your computer with your user’s permissions (files, network, apps).
  • The registry lists no key to fill in.

You need

  • npmNode.js installed (for the npx command)

Which app do you use?

Pick your app above. We remember it, so next time the button says “Install in …”.

Open Claude Desktop → Settings → Developer → Edit Config, and paste this into claude_desktop_config.json

{
  "mcpServers": {
    "mozak": {
      "command": "npx",
      "args": [
        "-y",
        "mozak",
        "mozak-mcp",
        "--config"
      ]
    }
  }
}

Run this in your terminal

claude mcp add mozak -- npx -y mozak mozak-mcp --config

This one needs values on its command line, so copy the settings and fill them in.

Or copy the settings yourself

Paste into .cursor/mcp.json in your project, or ~/.cursor/mcp.json

{
  "mcpServers": {
    "mozak": {
      "command": "npx",
      "args": [
        "-y",
        "mozak",
        "mozak-mcp",
        "--config"
      ]
    }
  }
}

Your browser asks to open the app, then the app shows the settings and asks you to confirm. Key values are left empty for you to fill in.

This one needs values on its command line, so copy the settings and fill them in.

Or copy the settings yourself

Paste into .vscode/mcp.json in your project

{
  "servers": {
    "mozak": {
      "type": "stdio",
      "command": "npx",
      "args": [
        "-y",
        "mozak",
        "mozak-mcp",
        "--config"
      ]
    }
  }
}

Your browser asks to open the app, then the app shows the settings and asks you to confirm. Key values are left empty for you to fill in.

This one runs on your computer, and ChatGPT can only connect to servers on the internet. Pick another app.

Your choice stays on this device in a cookie; no account needed.

These settings are generated from the maker’s registry entry. If the README says otherwise, follow the README.

What it can do

Runs on your computer; the tool list appears in your AI app after you install it.

Basic checksi

We check information anyone can see; we have not read the code. Checks that do not apply are left out of the count. · Checked Nov 7, 2026

  • ✓ Pass

    Maker matches the repo

    The registry name was verified with the GitHub account pitfa19, which owns the repo

  • … Not checked yet

    Still maintained

    We have not read this repo from GitHub yet

  • … Not checked yet

    Has a licence

    We have not read this repo from GitHub yet

  • – Not applicable

    Endpoint responds

    This one runs on your computer; there is no internet endpoint

  • – Not applicable

    Sign-in with OAuth

    Runs on your computer and uses local settings instead

  • … Not checked yet

    No hidden instructions in tool descriptions

    Runs on your computer; we can only read tool lists from internet endpoints

Compared with others in Docs, sheets & office worki

  • More stars than 0% of the other 2181 servers in this category
  • Published to the registry more recently than 67% of the other 2181 servers in this category
  • Basic checks equal to or better than 100% of the other 2181 servers in this category

Compared with every active server in this category in the registry. Stars count only when the publisher owns the repo.

Common questions

Is mozak safe?

It passed 1 of 1 applicable basic checks. We have not read the code. Before adding a key or token, see what access the server asks for and use a key with only the permissions you need.

What keys or codes does it need?

The registry does not list a required key, though some servers ask you to sign in on first use.

Does it run on my computer or in the cloud?

It runs on your computer. Your AI app starts it when needed, and data does not pass through the maker’s servers (unless it calls an online service itself).

Is it still maintained?

The GitHub repo last changed –, and the latest version (0.4.5) reached the registry 16 days ago.

Which apps can use it?

Apps that support MCP, such as Claude Desktop, Claude Code, Cursor and VS Code. You also need Node.js on your computer.

What can I use instead?

In the same category: Agent-Native Calendar, Excel MCP Server, MCP Server for Excel

Alternatives

Servers in the same category from other makers, sorted by stars.

Reading a card:★ GitHub starsiThree pillars: reviews · checks · usageiUpdated = latest version published to the registryi

Agent-Native Google Calendar - manage events, sync, and public booking

Runs in the cloud★ 7,620MITUpdated 2 months ago
Install

Excel MCP Server

by haris-musa

Lets AI create and edit Excel files itself, including formulas, charts and PivotTables, without Excel installed.

Runs on your computer★ 4,276MITUpdated 6 days ago
Install

Excel automation for AI - Sheets, Power Query, DAX, VBA, Tables, Ranges and more. Windows only.

Runs on your computer★ 847MITUpdated 3 days ago
Install

README from the maker

Open the README

Copied from the maker’s repo as written. PopNx MCP has not reviewed or endorsed it. HTML, images and scripts are removed. Links lead off this site.

One approved plan per project, shared by your coding agents, visible in Obsidian.

You work in ideas and plans, not code. Several agents share one plan without overwriting each other, and you follow their progress from your notes.

What it is · Example · See it run · Install · How it works · Obsidian · Safety · Tested · Docs


The problem

Coding agents forget the plan between sessions, so you explain the project again. When several agents work at once, one can overwrite what another just wrote. And an agent can start building before you agreed on what to build.

mozak fixes this with plain files:

  • Every agent reads the same plan. mozak.json holds it and mozak.md is a readable copy. You never explain the project twice.
  • Agents propose, you decide. Agreeing on *what* to do and agreeing to *start building* are two separate yeses.
  • Stale writes are refused. An agent that writes from an old copy gets stale-input and must reread. Nothing is overwritten.
  • One view across all your projects. The Obsidian plugin shows every project's plans and progress in one sidebar, with a Focus tab for the plans you starred.

No cloud, no account, no background service.

npm install -g mozak && mozak doctor --fix

A short example

Two agents read the same plan, then both propose at once. The first write lands. The second is refused because its copy is now old, so it cannot overwrite the first. This is real output: node scripts/readme-example.mjs reproduces it on a throwaway scope after npm run build.

$ mozak plan propose --scope my-app --request a.json   # agent A proposes

status: ok committed: true (exit 0)

$ mozak plan propose --scope my-app --request b.json # agent B proposes from the same, now old, snapshot status: stale-input committed: false stale-input: snapshot changed; reread before submitting a new operation: take the new snapshot from context as expected and use a new operationId (exit 3)

$ mozak plans --scope my-app # what is in the plan now awaiting (1)

  • Add dark mode [dark-mode] review: direction approval outcome: The app has a dark theme

The plan is now waiting for your yes. Agent B rereads and proposes again from the new snapshot.

See it run

The preview plays at double speed. In 76 seconds: you ask your agent in plain words, it reads the project and asks one question, drafts a plan and waits for your yes, records the test result, and the roadmap in Obsidian moves forward. The Obsidian side is a live capture of the plugin redrawing from the files. The terminal side is a rendered reenactment of the agent chat, with the same mozak commands run on disk. Click it for the full video with sound.

What it is

mozak gives every project one shared plan, stored as plain files in a folder of your Obsidian vault. It does two jobs:

  • Agents read the same plan. Agents that load the mozak skill or MCP server start from the same plan, so you don't have to paste the context into each one. Agents propose changes, but nothing becomes the plan until you approve it.
  • You follow it from Obsidian. Every project's plan shows up in your vault with a roadmap and progress, so parallel work is visible in one place. You approve the next step. An outcome turns green when an agent records passing evidence, such as a test run it reports. mozak does not run the tests itself, and outcomes you mark as yours also need your sign-off.

Why you can trust it:

  • It is just files. mozak.json holds the plan, mozak.md is a readable copy. They sit next to the project's README.md in your vault, and if the vault is in git, git keeps the history. No cloud, no account, no daemon.
  • You stay in charge. Agreeing on *what* to do and agreeing to *start building* are two separate approvals.
  • Many agents, no mess. A write made through mozak from an out-of-date copy is refused instead of overwriting someone else's work.
  • Your notes, not your code repos. Plans live only in your vaults, so a public code repo never carries them. A project can link to its code repo for one-click access, and mozak never fetches it.

It is for developers and researchers who already use coding agents and want them to share one plan. It works with any agent that can load a skill or an MCP server. Checked so far: jcode (skill and MCP) and Claude Code (MCP).

Install

You need Node 24.11 or newer (Node 26 works) and Git. Linux and macOS.

npm install -g mozak

mozak doctor --fix # links the mozak skill into ~/.claude/skills, ~/.jcode/skills or ~/.codex/skills if they exist

The npm package bundles the CLI and the MCP server in one package with no dependencies and no install scripts. Update with npm install -g mozak@latest.

Connect your agent to the MCP server (use your own inventory path):

# Claude Code

claude mcp add mozak -- mozak-mcp --config "$HOME/notes/mozak-inventory.json"

Codex

codex mcp add mozak -- mozak-mcp --config "$HOME/notes/mozak-inventory.json"

Any other MCP client: command mozak-mcp, arguments --config /path/to/mozak-inventory.json.

mozak scopes and the Obsidian sidebar tell you when a new release is out (MOZAK_NO_UPDATE_CHECK=1 turns that off). Changes are listed in CHANGELOG.md.

Install from source instead (portable install, no npm registry)

git clone https://github.com/pitfa19/mozak.git && cd mozak

npm ci --ignore-scripts && npm run build mkdir -p "$HOME/.local/share" node scripts/install.mjs --prefix "$(cd "$HOME/.local/share" && pwd -P)/mozak" --json "$HOME/.local/share/mozak/bin/mozak" --help

The installer copies a self-contained build into that folder. It uses no sudo, makes no network calls and does not touch your shell profile. Add --check to preview first. Two commands keep a source install working:

mozak upgrade --vault ~/Documents/notes   # newest tagged release; the old install is kept, plugin updated in that vault

mozak doctor --fix # checks the whole setup; fixes links, wrappers and plugin copies, prints the rest

Details, exit codes, recovery and known limits: docs/installation.md.

How it works

1. List your projects. One mozak-inventory.json names every project mozak may look at. Scope commands only touch projects listed there. Keep it at the top of your notes folder, so the Obsidian plugin finds it on its own.

{

"format": "mozak.local/v3", "scopes": [ { "id": "lab-notes", "root": "./research/lab-notes" }, { "id": "my-oss-tool", "root": "./creations/my-oss-tool", "repo": "github:you/my-oss-tool" } ] }

Every root is a folder inside an Obsidian vault (a folder with .obsidian at or above it). mozak refuses a scope outside a vault with scope-outside-vault. The optional repo is a display-only link to the project's code: the sidebar and mozak scopes show it, and mozak never clones, fetches or pushes it. Agents work in the code repo and name the project with --scope.

Coming from 0.5 with GitHub repo scopes or plans in a code repo? See Moving to 0.6.0: mozak scope move copies a scope into a vault folder and points its entry there.

2. Start a project. Give the folder a README.md and an AGENTS.md, then ask your agent to set it up with the mozak skill, or run mozak scope init yourself (see the skill's references/bootstrap.md). This creates the project's mozak.json and mozak.md. Creating a project is an owner command: agents cannot do it over MCP.

3. Agree on a plan.

  1. Propose. An agent drafts a plan. It waits as *awaiting approval*.
  2. Approve the direction. You say yes to *what*. This starts planning, not coding.
  3. Approve the build. A spec is attached and you approve implementation separately.
  4. Remember. Decisions and evidence are saved into the project's files, so the next session can read where the last one stopped.

4. Let agents find their way. mozak scopes lists every project with its plan counts. Agents get the same list from the MCP server.

See everything in Obsidian

The Obsidian plugin in packages/obsidian shows every project from your inventory in the sidebar, notes and GitHub repos alike. Tabs: Open, Focus (plans you starred) and Done. Open plans show a small progress strip, one segment per outcome. Click a plan to open its section in the scope's mozak.md, which reads like a note: purpose, a roadmap of the outcomes with the current stop, a status table and the full plan text. The sidebar is for reading and navigating. You approve plans in your agent chat, not in Obsidian. The plugin never approves or changes a plan and never writes plan state. It may regenerate a stale mozak.md, the same as mozak render, and the star on a plan row edits that scope's focus list in mozak-inventory.json (the same as mozak focus add|remove).

Install it from Obsidian: Settings, Community plugins, Browse, search for mozak, then Install and Enable (directory page). From a source checkout you can also run:

node scripts/install-obsidian-plugin.mjs /path/to/your/vault --build

The plugin finds mozak-inventory.json by searching upward from the vault folder, or you can set its path in the plugin settings. A project with a repo link shows a link icon that opens its code repo in the browser.

Plugin disclosures (desktop-only, `isDesktopOnly: true`): When your mozak-inventory.json lists scopes outside the current vault (in your other vaults), the plugin reads their mozak.json, mozak.md, mozak.annotations.json (when present and annotations are on for that scope), README.md and AGENTS.md files from those configured paths and watches them when Auto-refresh is enabled. It never scans folders not listed in your inventory and never approves or changes a plan (it may regenerate a stale mozak.md projection when there are no unmerged hand edits, and the star toggle edits the focus lists in mozak-inventory.json). The only network use is the Sync vaults button, the same as mozak sync: your local git with your existing git login runs git fetch in each vault repository that holds a scope, then git merge --ff-only under the scope locks. It never pushes and never merges diverged histories. Nothing runs in the background. The plugin never contacts GitHub for a scope's repo link (the link is opened by your browser only when you

Only the beginning is shown; read the rest in the repo

Ask PopNx