Employee Leave Manager MCP
Enables managing employee leave through natural language, including checking balances, applying for or canceling leave, and viewing history, with data stored in a local SQLite database.
README
Employee Leave Manager — MCP
An MCP server that exposes employee leave management (balances, applying for leave, history, etc.) as tools that Claude Desktop (or any MCP client) can call. Data is stored in a local SQLite database.
Prerequisites
- uv — Python dependency manager
- Claude Desktop
Setup
Clone the repo:
git clone https://github.com/TharushiDinushika/Employee-Leave-Manager-MCP.git
cd Employee-Leave-Manager-MCP
Create the project and install dependencies:
uv sync
Activate Virtual Environment
.venv\Scripts\activate
Initialize the database
Creates the SQLite schema and seeds it with sample employees, leave types, and balances:
python init_db.py
Options:
python init_db.py --reset # drop all tables and reseed from scratch
python init_db.py --no-seed # create schema only, skip sample data
Run the server
uv run mcp dev server.py
This opens the MCP Inspector, where you can list and test the available tools directly before connecting a client.
Connect to Claude Desktop
Open Claude Desktop and check whether the connector appears automatically.

If it doesn't show up, add it manually:
- Go to File → Settings → Developer → Edit Config
- Add the following block to the config file (update the path to match where you cloned the project):
{
"mcpServers": {
"EmployeeLeaveManager": {
"command": "uv",
"args": [
"--directory",
"C:\\path\\to\\Employee-Leave-Manager-MCP",
"run",
"server.py"
]
}
}
}
- Restart Claude Desktop.
- Open a new chat and try asking Claude about employee leave — e.g. "What's E001's leave balance?"
Functionality
The server exposes the following tools:
| Tool | Description |
|---|---|
get_leave_balance(employee_id, leave_type) |
Check remaining leave balance for an employee. Returns balance for a specific leave type, or all types if omitted. |
apply_leave(employee_id, leave_dates, leave_type="annual", reason) |
Apply for leave on one or more specific dates. Validates date format, rejects past or duplicate dates, and checks balance before deducting. |
cancel_leave(employee_id, leave_dates) |
Cancel previously approved leave and refund the balance. |
get_leave_history(employee_id, leave_type) |
View an employee's leave history, optionally filtered by leave type. |
get_leave_types() |
List all available leave types (annual, sick, casual, unpaid) and their default annual allotment. |
list_employees(department) |
List all employees, optionally filtered by department. |
add_employee(employee_id, name, department, email, manager_id) |
Onboard a new employee with default leave balances. |
get_employee_details(employee_id) |
Get an employee's profile (name, department, email, manager). |
It also exposes two resources:
greeting://{name}— a personalized greetingemployee://{employee_id}/summary— a quick profile + balance summary for an employee
Data model
Data lives in a local SQLite database (leave_manager.db), with four tables:
employees, leave_types, leave_balances, and leave_requests. Schema
creation and seeding are handled separately by init_db.py, so the server
itself never modifies the schema — it just reads/writes data.
Project structure
mcp-server/
├── server.py # MCP server: tools, resources, DB queries
├── init_db.py # Database schema creation + seeding (run once)
├── leave_manager.db # SQLite database (created after running init_db.py)
推荐服务器
Baidu Map
百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。
Playwright MCP Server
一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。
Magic Component Platform (MCP)
一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。
Audiense Insights MCP Server
通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。
VeyraX
一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。
graphlit-mcp-server
模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。
Kagi MCP Server
一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。
e2b-mcp-server
使用 MCP 通过 e2b 运行代码。
Neon MCP Server
用于与 Neon 管理 API 和数据库交互的 MCP 服务器
Exa MCP Server
模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。