LimeAI MCP Server
README.md
# LimeAI MCP Server (Python)
### 环境要求
- Python 3.10 或更高版本。
## 1. 设置环境
首先,我们来安装`uv`:
MacOS/Linux
```bash MacOS/Linux
curl -LsSf https://astral.sh/uv/install.sh | sh
```
Windows
```powershell Windows
powershell -ExecutionPolicy ByPass -c "irm https://astral.sh/uv/install.ps1 | iex"
```
## 2. 获取 LimeAI MCP Server
```bash
# LimeAI Mcp Server 官方开源仓库下载
git@github.com:liuwenlonghub/limeai-mcp-server.git
cd limeai-mcp-server
# 创建虚拟环境并激活
uv venv
source .venv/bin/activate
# 安装依赖
uv add "mcp[cli]" httpx
```
#### 获取 `LIMEAI_ACCESS_TOKEN`
- 登录 LimeAI(www.limeai.net) > 个人资料 > 访问令牌
## 3. 在Cursor中使用
`Cursor -> Preferences -> Cursor Settings -> MCP -> Add new global MCP`
添加以下配置内容(其它MCP客户端类似):
```json
{
"mcpServers": {
"limeai-mcp-server": {
"command": "uv",
"args": [
"--directory",
"{YOUR_PATH}/limeai-mcp-server/",
"run",
"main.py"
],
"env": {
"LIMEAI_ACCESS_TOKEN": "<USER_LIMEAI_ACCESS_TOKEN>"
}
}
}
}
```
回到配置,此时 LimeAI MCP Server已经启用
TDQS
A3.5/5.0
Scored across 3 tools
Disambiguation5/5
Each tool has a clearly distinct purpose: retrieving user info, saving markdown, and saving HTML. There is no overlap or ambiguity between them.
Naming Consistency5/5
All tool names follow a consistent verb_noun pattern using snake_case: get_user_info, save_markdown, save_html. The pattern is predictable and uniform.
Tool Count5/5
With just 3 tools, the server is tightly scoped to its apparent purpose of saving content in markdown/HTML and retrieving user info. No unnecessary tools are present, and the count feels appropriate.
Completeness4/5
Core save workflows for both markdown and HTML are covered, along with user info retrieval. However, there are no tools for list, get, update, or delete saved content, which is a minor gap that agents might need to work around.
Maintenance
ActivityInactive
ResponsivenessNo issues