Frappe/ERPNext MCP Server
Provides AI assistance for Frappe/ERPNext development with tools for creating DocTypes, running bench commands, and managing apps.
README
Frappe/ERPNext MCP Server
A Model Context Protocol (MCP) server that provides AI assistance for Frappe/ERPNext development. This server offers tools to help with creating DocTypes, running bench commands, managing apps, and other Frappe development tasks.
Features
- DocType Creation: Automatically generate DocType JSON definitions and Python controllers
- Bench Commands: Execute bench commands for development workflows
- App Management: Create, install, and manage Frappe apps
- API Endpoints: Generate custom API endpoints
- Database Operations: Run migrations and database management
- App Structure Analysis: Get detailed app directory structures
Installation
-
Clone or download this repository
-
Install dependencies:
npm install -
Build the server:
npm run build
Configuration
Set the FRAPPE_PATH environment variable to point to your Frappe bench directory:
export FRAPPE_PATH="/path/to/your/frappe/bench"
If not set, it defaults to /Users/mac/ERPNext/mkahawa.
Usage
With GitHub Copilot Chat (SSE Transport)
The server now supports SSE transport for use with GitHub Copilot Chat:
-
Build and run the SSE server:
npm run build npm run sse -
The server will be automatically discovered by GitHub Copilot Chat at
http://localhost:3000/sse
With Claude Desktop (Stdio Transport)
Add the following configuration to your Claude Desktop config file (~/Library/Application Support/Claude/claude_desktop_config.json on macOS):
{
"mcpServers": {
"frappe-dev": {
"command": "node",
"args": ["/Users/mac/ERPNext/mkahawa/frappe-mcp-server/dist/index.js"],
"env": {
"FRAPPE_PATH": "/Users/mac/ERPNext/mkahawa"
}
}
}
}
With other MCP clients
Run the server directly:
node dist/index.js
Available Tools
frappe_create_doctype
Creates a new Frappe DocType with JSON definition and Python controller.
Parameters:
app_name: Name of the Frappe appdoctype_name: Name of the DocTypemodule: Module where DocType belongsfields: Array of field definitionsis_submittable: Whether the DocType is submittable (default: false)is_child: Whether this is a child DocType (default: false)
frappe_run_bench_command
Executes bench commands for Frappe development.
Parameters:
command: Bench command to executesite: Site name (optional)cwd: Working directory (optional)
frappe_get_app_structure
Gets the directory structure of a Frappe app.
Parameters:
app_name: Name of the Frappe app
frappe_create_api_endpoint
Creates a custom API endpoint for a Frappe app.
Parameters:
app_name: Name of the Frappe appendpoint_name: Name of the API endpointmethod: HTTP method (get, post, put, delete) (default: get)code: Python code for the API endpoint
frappe_migrate_database
Runs database migration for a Frappe site.
Parameters:
site: Site name to migrate
frappe_install_app
Installs a Frappe app on a site.
Parameters:
app_name: Name of the app to installsite: Site name
frappe_create_app
Creates a new Frappe app.
Parameters:
app_name: Name of the new apptitle: Title of the apppublisher: Publisher namedescription: App description (optional)
Example Usage
Creating a DocType
{
"app_name": "my_app",
"doctype_name": "Customer Order",
"module": "Orders",
"fields": [
{
"fieldname": "customer",
"label": "Customer",
"fieldtype": "Link",
"options": "Customer",
"reqd": true
},
{
"fieldname": "order_date",
"label": "Order Date",
"fieldtype": "Date",
"reqd": true
},
{
"fieldname": "total_amount",
"label": "Total Amount",
"fieldtype": "Currency",
"reqd": true
}
]
}
Running Bench Commands
{
"command": "migrate",
"site": "my_site.local"
}
Development
To modify the server:
- Edit files in the
src/directory - Run
npm run buildto compile TypeScript - Test with
npm run devfor development mode
Requirements
- Node.js 18+
- Frappe/ERPNext bench environment
- Access to Frappe bench commands
License
MIT License - see LICENSE file for details
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。