ERPNext MCP Server
Enables interaction with ERPNext instances via its REST API to manage documents, inventory, and reports. It supports full CRUD operations, submittable document workflows, and schema inspection through natural language.
README
ERPNext MCP Server
MCP (Model Context Protocol) server for ERPNext REST API, built with FastMCP and Python.
Features
- CRUD — List, get, create, update, delete documents
- Workflow — Submit and cancel submittable documents
- Reports — Run ERPNext query reports
- Schema — Inspect DocType field definitions, list all DocTypes
- Inventory — Stock balance, stock ledger, item prices
- Trading — Document conversion (e.g. Quotation → Sales Order), party balance
- Supplier/Customer — Get complete details with address, phone, contacts; supports alias search
- Files — Upload, list, download files
- Helpers — Link search (autocomplete), document count, generic method calls
Requirements
- Python >= 3.11
- uv (recommended) or pip
- ERPNext instance with API key/secret
Setup
# Clone the repo
git clone <repo-url> && cd erpnext-mcp
# Create .env file
cat > .env << 'EOF'
ERPNEXT_URL=https://your-erpnext-instance.com
ERPNEXT_API_KEY=your_api_key
ERPNEXT_API_SECRET=your_api_secret
EOF
# Install dependencies
uv sync
Run
set -a && source .env && set +a && uv run erpnext-mcp
Available Tools
| Tool | Description |
|---|---|
list_documents |
List documents with filters, sorting, pagination |
get_document |
Get a single document by name |
create_document |
Create a new document |
update_document |
Update an existing document |
delete_document |
Delete a document |
submit_document |
Submit a submittable document |
cancel_document |
Cancel a submitted document |
run_report |
Execute an ERPNext report |
get_count |
Get document count with optional filters |
get_list_with_summary |
List documents with total count |
run_method |
Call any whitelisted server-side method |
search_link |
Link field autocomplete search |
list_doctypes |
List all available DocType names |
get_doctype_meta |
Get field definitions for a DocType |
get_stock_balance |
Real-time stock balance from Bin |
get_stock_ledger |
Stock ledger entries (inventory history) |
get_item_price |
Item prices from price lists |
make_mapped_doc |
Document conversion (e.g. SO → DN) |
get_party_balance |
Outstanding balance for Customer/Supplier |
get_supplier_details |
Get supplier with address, phone, contacts (supports alias search) |
get_customer_details |
Get customer with address, phone, contacts (supports alias search) |
upload_file |
Upload a local file to ERPNext (by file path) |
upload_file_from_url |
Upload a file from URL |
list_files |
List files attached to a document |
download_file |
Download a file by URL |
get_file_url |
Get download URL for a file |
MCP Client Configuration
Add to your MCP client config (e.g. Claude Desktop claude_desktop_config.json):
{
"mcpServers": {
"erpnext": {
"command": "uv",
"args": ["--directory", "/path/to/erpnext-mcp", "run", "erpnext-mcp"],
"env": {
"ERPNEXT_URL": "https://your-erpnext-instance.com",
"ERPNEXT_API_KEY": "your_api_key",
"ERPNEXT_API_SECRET": "your_api_secret"
}
}
}
}
Project Structure
src/erpnext_mcp/
├── server.py # MCP tool definitions (FastMCP)
├── client.py # ERPNext REST API client (httpx async)
└── types.py # Pydantic models
License
MIT
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。