Metadata-Version: 2.4
Name: agentfenster-mcp
Version: 0.1.0
Summary: Launcher for the agentfenster MCP server - Windows GUI automation on a hidden second desktop.
Author-email: agentfenster <hello@agentfenster.com>
License: Proprietary
Project-URL: Homepage, https://agentfenster.com
Project-URL: Documentation, https://agentfenster.com/mcp
Project-URL: Repository, https://github.com/clemensjl/agentfenster-mcp
Project-URL: Issues, https://github.com/clemensjl/agentfenster-mcp/issues
Project-URL: Discord, https://discord.gg/J9MfKP255Y
Keywords: mcp,model-context-protocol,claude,claude-code,windows,automation,gui-automation,computer-use,uiautomation
Classifier: Development Status :: 4 - Beta
Classifier: Environment :: Console
Classifier: Environment :: Win32 (MS Windows)
Classifier: Intended Audience :: Developers
Classifier: Intended Audience :: End Users/Desktop
Classifier: License :: Other/Proprietary License
Classifier: Operating System :: Microsoft :: Windows :: Windows 10
Classifier: Operating System :: Microsoft :: Windows :: Windows 11
Classifier: Programming Language :: Python :: 3.11
Classifier: Programming Language :: Python :: 3.12
Classifier: Programming Language :: Python :: 3.13
Classifier: Topic :: Desktop Environment
Classifier: Topic :: Software Development :: Testing
Classifier: Topic :: Utilities
Requires-Python: >=3.11
Description-Content-Type: text/markdown

# agentfenster-mcp

`agentfenster-mcp` is the launcher for the agentfenster MCP server. The server
lets Claude Code and other MCP clients drive real Windows applications on a
hidden second desktop inside your own session — with the same logins, cookies
and VPN — while your screen, mouse and keyboard stay yours.

Every action is verified against the window state afterwards. A click that
lands nowhere is reported as a failure, never as a silent success.

Full documentation: <https://agentfenster.com/mcp>

mcp-name: com.agentfenster/agentfenster

<!-- mcp-name: io.github.clemensjl/agentfenster -->

## What this package is, and what it is not

This package contains **no product code**. It is the small public launcher that
makes the documented one-liner work:

```
claude mcp add agentfenster -- agentfenster-mcp
```

The MCP server itself ships with the **agentfenster app**, a commercial,
closed-source Windows program distributed from <https://agentfenster.com>.

- If the app is installed, `agentfenster-mcp` hands over to the real server and
  gets out of the way.
- If it is not, `agentfenster-mcp` prints what is missing and how to fix it,
  then exits with status 1 — instead of leaving your MCP client with a bare
  `command not found`.

Installing this package alone does not give you the server.

## Install

1. Install agentfenster on Windows 10 or 11 from <https://agentfenster.com>.
   The MCP server is part of the app. Every plan starts with a 3-day trial, no
   card and no account required.
2. Register the server with Claude Code:

   ```
   claude mcp add agentfenster -- agentfenster-mcp
   ```

   Any MCP client that can launch a local stdio server works the same way.

Then just ask:

```
> open the character map on the agent desktop and turn on the advanced view
```

## The 11 tools

| Tool | What it does |
|------|--------------|
| `list_windows` | Lists the windows this machine can drive, with title, class, minimized state and whether a picture can be captured. |
| `desktop_task` | Runs a full GUI task in one window and returns a single paragraph: what happened, whether it worked, and if not, which step failed and why. |
| `inspect_window` | Returns the capped, filtered element tree of one window as a compact list, for planning clicks yourself. |
| `replay` | Lists recent runs, or returns one run's step-by-step log, with a link to the visual replay in the app. |
| `open_app` | Launches a program on the shared agent desktop and reports its first titled window. |
| `read_desktop` | Reads the compact element tree of a window on the agent desktop, one line per interactive element, no screenshot. |
| `click` | Clicks one element, then verifies against the desktop state and states explicitly whether anything changed. |
| `type_text` | Types text into one element on the agent desktop, then verifies whether anything changed. |
| `press_key` | Presses a key or combination such as `ctrl+s`, `enter`, `tab` or `alt+f4`, with verification afterwards. |
| `scroll` | Scrolls a window or pane and reports whether the content actually moved. |
| `agent_windows` | Lists the windows currently open on the shared agent desktop, frontmost first. |

## How it works

- The agent works on a second Win32 desktop inside your session. It never takes
  your foreground, mouse or keyboard, and you can watch it live in the app.
- Tools read the UI Automation element tree instead of screenshots: one line per
  interactive element, capped at 200 elements per window.
- The click loop of `desktop_task` runs inside the server on a model endpoint
  you configure, so its retries never enter your context.
- Risky actions such as delete, send, pay or install are denied by default on
  MCP calls and reported, instead of blocking on a confirmation nobody can
  answer.
- agentfenster sends no telemetry. The only network traffic is the model
  endpoint you configure for the click loop.

## Requirements

| Aspect | Details |
|--------|---------|
| Platform | Windows 10 or 11 |
| Transport | stdio, local |
| Runtime | Python 3.11 or newer |
| Prerequisite | the agentfenster app (commercial, 3-day trial) |

## Links

- Product and download: <https://agentfenster.com>
- MCP documentation: <https://agentfenster.com/mcp>
- Pricing: <https://agentfenster.com/#pricing>
- Documentation and issue tracker: <https://github.com/clemensjl/agentfenster-mcp>
- Discord: <https://discord.gg/J9MfKP255Y>
- Support: hello@agentfenster.com
- Privacy: <https://agentfenster.com/privacy>
- Terms: <https://agentfenster.com/terms>

## Rights

agentfenster is commercial, closed-source software. No open-source license is
granted here. This launcher package is published so that the documented install
command works and so that MCP directories can resolve it. Use of the
agentfenster app itself is governed by the terms at
<https://agentfenster.com/terms>.
