Skip to main content
Glama

GPT Plus Plus

Let MCP-enabled GPT web clients or AI clients control local development tools on your Windows PC.

Note: The original source code has been deleted by GPT and cannot be recovered. This source is a snapshot; the EXE version is recommended.

GPT Plus Plus is a Windows desktop program. It can provide the local project directory you choose to GPT or other MCP-enabled clients, letting the AI act under permission control to:

  • View project files and directories;

  • Search code;

  • Modify files;

  • Run tests and development commands;

  • View Git status, diffs, and commit history;

  • View images and other project resources.

The default permission is safe. No need to install Docker or WSL.

Note: The GPT web client must support custom MCP services or remote connection features. The interface names may differ across accounts, regions, and client versions.

The simplest way: Windows EXE

Download it from the GitHub Releases page:

GPT++-0.5.0.exe

Double-click to run; no separate installation of uvx, Python, MCP CLI, or cloudflared is required.

Usage steps

  1. Open GPT++-0.5.0.exe.

  2. In the left pane, select or create a workspace pointing to the project folder you want the AI to operate on.

  3. Keep the permission mode set to safe.

  4. Start the local MCP service.

  5. If using it only on this machine, use the displayed local address:

    http://127.0.0.1:<端口>/mcp
  6. To let the GPT web client connect from the public internet, choose Quick Tunnel or Fixed Domain Tunnel under "Public Access".

  7. Wait until the interface shows the public address is ready, then copy the displayed MCP address:

    https://<公网域名>/mcp
  8. In an MCP-enabled GPT web client or AI client, add a remote MCP service and enter the above address as the connection address.

  9. Complete OAuth or Bearer Token authentication as prompted; a configuration without authentication is suitable only for local testing and should not be exposed to the public internet.

  10. Once connected, you can directly tell GPT:

    请读取当前项目结构,运行测试,并告诉我第一个失败的原因。

Two Tunnel modes

  • Quick Tunnel: No Tunnel Token needs to be filled in; a temporary public address is generated. The address may change on each restart.

  • Fixed Domain Tunnel: You need to enter your own Tunnel Token and public address in the desktop settings. Credentials are stored only in the local configuration and are never written into the source code or the EXE.

The program will not show "Ready" merely because cloudflared printed a URL; it also checks the public MCP metadata and the /mcp response, avoiding the situation where "it looks connected but GPT cannot call the tools."

Related MCP server: COP Plug

How the GPT web client controls local tools

The connection relationship is as follows:

GPT 网页端
    ↓ HTTPS + MCP
公网 Tunnel 地址 /mcp
    ↓
GPT Plus Plus 本地服务
    ↓
safe 权限控制
    ↓
你选择的 Windows 项目文件夹

GPT cannot bypass the MCP service to read your computer directly. It can only call the tools exposed by the service; the service checks paths, permissions, commands, and workspace boundaries.

It is recommended to connect only one test project the first time and keep safe mode. Do not use folders containing passwords, private keys, browser profiles, or other sensitive information as a workspace.

Local development and running

If you want to run from source instead of using the EXE:

cd /d D:\测试文件夹\gpt-plus-plus
python -m venv .venv
.venv\Scripts\python.exe -m pip install -e ".[desktop,dev]"
.venv\Scripts\gpt-plus-plus-desktop.exe

You can also use:

python -m pip install -e ".[desktop,dev]"
gpt-plus-plus-desktop

Run tests:

python -m unittest discover -s tests -p "test_*.py" -t . -v

Safety notes

  • safe is used by default, but it is not a complete sandbox like Docker.

  • Do not commit Tunnel Tokens, Bearer Tokens, OAuth passwords, or private keys to GitHub.

  • Do not upload .env, profiles, OAuth state, logs, build, dist, or the EXE to the source repository.

  • Remote connections must use HTTPS and authentication.

  • Only set as a workspace the project directories you are willing to let the AI operate on.

  • Do not casually use trusted or dangerous mode in untrusted projects.

Source code and packaging

The source repository does not contain cloudflared.exe. The Windows Release workflow downloads a pinned version of the official cloudflared, checks the SHA-256 and Authenticode signatures, and then packages.

Windows packaging instructions:

packaging/windows/BUILDING.md

Documentation

License

This project is licensed under Apache License 2.0. Attribution and third-party notices are in NOTICE and THIRD_PARTY_NOTICES.md.

A
license - permissive license
Not graded
quality - not tested
B
maintenance

Maintenance

Maintainers
Response time
Release cycle
1Releases (12mo)
Commit activity

Resources

Unclaimed servers have limited discoverability.

Looking for Admin?

If you are the server author, to access and configure the admin panel.

Related MCP Servers

  • A
    license
    Not graded
    quality
    C
    maintenance
    Enables AI clients to perform local code search, indexing, and analysis across Java, JavaScript/TypeScript, .NET/C#, and Python projects through the MCP protocol.
    2
    Apache 2.0
  • F
    license
    Not graded
    quality
    A
    maintenance
    Enables AI assistants to execute PowerShell commands, manage files, inspect projects, run Git operations, and monitor system information on Windows through a local MCP server.
  • A
    license
    Not graded
    quality
    C
    maintenance
    Enables AI agents to seamlessly integrate with the Windows operating system, performing tasks such as file navigation, application control, UI interaction, and QA testing via the MCP protocol.
    MIT
  • A
    license
    Not graded
    quality
    B
    maintenance
    Enables ChatGPT to securely operate a single Windows development workspace via a local MCP server, offering file editing, Git status, static analysis, approved test/build, and limited ADB operations with audit logging.
    MIT

View all related MCP servers

Related MCP Connectors

  • Persistent memory and cross-session learning for AI coding assistants (hosted remote MCP).

  • Personal assistant MCP server with search, execute, packages, jobs, secrets, and integrations.

  • An MCP server that gives your AI access to the source code and docs of all public github repos

View all MCP Connectors

Latest Blog Posts

MCP directory API

We provide all the information about MCP servers via our MCP API.

curl -X GET 'https://glama.ai/api/mcp/v1/servers/ranhunya/gpt-plus-plus'

If you have feedback or need assistance with the MCP directory API, please join our Discord server