Jelly Memo Local MCP Server

这是发布 Jelly Memo 用户设备专用 AI 协作服务器和 Windows 安装程序的仓库。 首先提供韩文指南，English documentation 在下方。

韩语

提供什么？

• 开源 Python MCP 服务器

• 服务器运行和测试脚本

• PyInstaller 可执行文件构建定义

• Inno Setup Windows 安装程序构建定义

• 可直接安装的 Windows 安装程序 Release

安装此工具后，支持 MCP 的 AI 工具可以查找 Jelly Memo 的备忘录并提出更改建议。是否应用更改以及自动保存设置由 Jelly Memo 应用管理。

本地专用·隐私保护

• 服务器仅在用户计算机的 127.0.0.1:8765 上运行。

• 拒绝绑定到 0.0.0.0 或外部网络地址。

• 服务器不直接连接 Firebase、Firestore 或 Firebase Storage。

• 备忘录快照和 AI 更改建议仅在 Jelly Memo 应用与此本地服务器之间交换。

• 可以选择性地设置配对令牌（pairing token）以限制快照更改请求。

此仓库不包含 Jelly Memo Flutter 应用本体、Firebase 配置、用户备忘录或认证信息。

最简单的安装方法

1. 打开最新 Release。

2. 下载 JellyMemoMcpInstaller-<version>-windows-x64.exe。

3. 运行安装程序，并在 Jelly Memo 设置的 AI 协作工具中确认连接状态。

当前安装程序未应用 Authenticode 代码签名，因此 Windows SmartScreen 可能会显示警告。在这种情况下，请直接确认下载的文件和此仓库的 Release 来源后再决定是否安装。

从源代码运行

要求：Windows、Python 3.10 或更高版本

git clone https://github.com/egm7126/simpleMemo-releases.git
cd simpleMemo-releases
py -3.10 -m venv .venv
.\.venv\Scripts\python.exe -m pip install -e .
.\scripts\run_companion.ps1 -PythonPath ".\.venv\Scripts\python.exe"

检查状态：

Invoke-RestMethod http://127.0.0.1:8765/health

测试和 Windows 安装程序构建

.\.venv\Scripts\python.exe -m unittest discover -s tests -v
.\.venv\Scripts\python.exe -m compileall jellymemo_mcp_companion tests
.\.venv\Scripts\python.exe -m pip install -e ".[build]"
.\packaging\build_server.ps1 -PythonPath ".\.venv\Scripts\python.exe"
.\packaging\build_installer.ps1

构建安装程序需要 Inno Setup 6。输出文件生成在 build/mcp_companion/ 下。

开源与贡献

此本地 MCP 服务器以 MIT License 发布。您可以使用、审查、修改和重新分发源代码。请通过 GitHub Issues 或 Pull Request 提交错误报告和改进建议。

Related MCP server: Bruin

English

This repository publishes the local AI collaboration server for Jelly Memo, together with its Windows installer releases.

What is included?

• Open-source Python MCP server

• Server run and test scripts

• PyInstaller executable build definition

• Inno Setup Windows installer definition

• Ready-to-install Windows installer releases

Once installed, MCP-compatible AI tools can find Jelly Memo notes and propose changes. Jelly Memo controls whether proposals are applied and saved automatically.

Local-only privacy model

• The server listens only on 127.0.0.1:8765 on the user's computer.

• Binding to 0.0.0.0 or another external network address is rejected.

• The server does not connect directly to Firebase, Firestore, or Firebase Storage.

• Memo snapshots and AI proposals are exchanged only between Jelly Memo and this local server.

• An optional pairing token can restrict snapshot mutation requests.

This repository does not contain the Jelly Memo Flutter application, Firebase configuration, user notes, or credentials.

Easiest installation

1. Open the latest Release.

2. Download JellyMemoMcpInstaller-<version>-windows-x64.exe.

3. Run the installer and check the connection under AI collaboration tools in Jelly Memo settings.

The installer is currently not Authenticode-signed, so Windows SmartScreen may display a warning. Verify the downloaded file and its GitHub Release source before choosing whether to install it.

Run from source

Requirements: Windows and Python 3.10 or newer

git clone https://github.com/egm7126/simpleMemo-releases.git
cd simpleMemo-releases
py -3.10 -m venv .venv
.\.venv\Scripts\python.exe -m pip install -e .
.\scripts\run_companion.ps1 -PythonPath ".\.venv\Scripts\python.exe"

Check server health:

Invoke-RestMethod http://127.0.0.1:8765/health

Test and build the Windows installer

.\.venv\Scripts\python.exe -m unittest discover -s tests -v
.\.venv\Scripts\python.exe -m compileall jellymemo_mcp_companion tests
.\.venv\Scripts\python.exe -m pip install -e ".[build]"
.\packaging\build_server.ps1 -PythonPath ".\.venv\Scripts\python.exe"
.\packaging\build_installer.ps1

Building the installer requires Inno Setup 6. Build outputs are written under build/mcp_companion/.

Open source and contributing

This local MCP server is available under the MIT License. You may use, inspect, modify, and redistribute the source. Bug reports and improvements are welcome through GitHub Issues and Pull Requests.