ctbcsec-api-mcp-server

ctbcsec-api-mcp-server

Enables LLMs to interact with CTBC Securities CTS Trading API for stock, futures, and options trading operations, including order placement, account management, and position queries.

Category
访问服务器

README

CTS Trading API MCP Server

A Model Context Protocol (MCP) server providing comprehensive access to CTBC Securities CTS Trading API. Enables LLMs like Claude to interact with the trading platform for stock, futures, and options trading operations.

Features

  • Stock Trading: Place, modify, and cancel stock orders
  • Futures/Options Trading: Full support for derivatives trading
  • Account Management: Query accounts, positions, orders, and matches
  • Type-Safe: Comprehensive Pydantic models for all operations
  • Event-Driven: Async response handling for query operations
  • MCP Resources: Access to configuration and connection status

Requirements

  • Operating System: Windows (COM object dependency)
  • Python: 3.10 or higher
  • CTS Trading API: Must be installed and registered on the system
  • Dependencies:
    • mcp >= 1.1.0
    • pywin32 >= 306
    • pydantic >= 2.0

Installation

  1. Clone or download the repository

  2. Install CTS Trading API (from CTBC Securities)

    • Ensure the DJTRADEOBJLibCTS.TradeApp COM object is registered
  3. Install Python dependencies

    # Using uv (recommended)
    uv pip install -e .
    
    # Or using pip
    pip install -e .
    
  4. Configure the server

    Edit appsetting.json with your trading server endpoint:

    {
      "TradeDas": "your.trading.server.com/tradedas"
    }
    

Usage

Testing with MCP Inspector

Test the server using the MCP development tools:

uv run mcp dev d:\ctbcsec-api-mcp-server\ctbcsec_mcp\server.py

This launches the MCP Inspector where you can:

  • View all available tools
  • Test tool schemas
  • Execute tools interactively
  • Monitor responses

Integration with Claude Desktop

Install the server to Claude Desktop:

uv run mcp install d:\ctbcsec-api-mcp-server\ctbcsec_mcp\server.py

Or manually add to Claude Desktop configuration (%APPDATA%\Claude\claude_desktop_config.json):

{
  "mcpServers": {
    "ctbcsec-trading": {
      "command": "uv",
      "args": [
        "--directory",
        "d:\\ctbcsec-api-mcp-server",
        "run",
        "ctbcsec-mcp"
      ]
    }
  }
}

Available Tools

Authentication & Connection

  • initialize: Initialize the CTS Trading API with server configuration
  • login: Authenticate user with trading system
  • connect: Connect to the trading server
  • disconnect: Disconnect from trading server
  • logout: Logout from trading system
  • get_accounts: Retrieve all available trading accounts
  • get_connection_status: Get current connection status
  • set_lot_size: Set lot size data for specific stocks

Stock Trading

  • stock_new_order: Place a new stock order
  • stock_modify_order: Modify an existing stock order
  • stock_cancel_order: Cancel an existing stock order
  • stock_query_order: Query stock orders
  • stock_query_match: Query stock trade matches
  • stock_query_position: Query stock positions

Futures/Options Trading

  • futopt_new_order: Place a new futures/options order
  • futopt_modify_order: Modify an existing futures/options order
  • futopt_cancel_order: Cancel an existing futures/options order
  • futopt_query_order: Query futures/options orders
  • futopt_query_match: Query futures/options matches
  • futopt_query_oi: Query futures/options open interest
  • futopt_query_equity: Query futures/options account equity

Resources

  • config://appsetting: Current server configuration
  • status://connection: Current connection and authentication status

Example Usage Flow

# 1. Initialize the trading API
initialize(trade_das_url="apsit.ectest.ctbcsec.com/tradedas")

# 2. Login
login(user_id="your_user_id", password="your_password")

# 3. Connect to trading server
connect()

# 4. Get available accounts
accounts = get_accounts()

# 5. Place a stock order
stock_new_order(
    account_id="1234567",
    stock_id="2330",
    quantity="1000",
    price="500",
    buy_sell=BuySell.BUY,
    price_type=PriceType.LIMIT
)

# 6. Query positions
stock_query_position(account_id="1234567")

# 7. Cleanup
disconnect()
logout(user_id="your_user_id")

Enumerations

TradeType (Stock)

  • REGULAR = 0: Regular trading
  • AFTER_HOURS_ODD_LOT = 1: After-hours odd lot
  • AFTER_HOURS = 2: After-hours
  • EMERGING = 5: Emerging stock
  • INTRADAY_ODD_LOT = 7: Intraday odd lot

OrderType (Stock)

  • CASH = 0: Cash order
  • MARGIN = 1: Margin trading
  • SHORT = 2: Short selling
  • DAY_TRADING_SELL_FIRST = 16: Day trading sell first

BuySell

  • BUY = 1: Buy order
  • SELL = 2: Sell order

PriceType

  • LIMIT = 0: Limit price
  • LIMIT_UP = 1: Limit up
  • LIMIT_DOWN = 2: Limit down
  • FLAT = 3: Flat
  • MARKET = 4: Market price

OrderCondition

  • ROD = 0: Rest of Day
  • IOC = 1: Immediate or Cancel
  • FOK = 2: Fill or Kill

ProductType (Futures/Options)

  • FUTURES = 0: Futures
  • OPTIONS = 1: Options
  • COMPLEX_OPTIONS = 2: Complex options
  • COMPLEX_FUTURES = 3: Complex futures

Architecture

Components

  • server.py: FastMCP server with tool definitions and lifespan management
  • models.py: Pydantic models for type-safe structured data
  • wrapper.py: COM object wrapper with event handling and thread safety
  • __init__.py: Package initialization

Data Flow

  1. MCP tool receives request with typed parameters
  2. Server validates input using Pydantic models
  3. Wrapper executes COM object method with thread safety
  4. Event handler queues async responses
  5. Structured response returned to LLM

Thread Safety

The wrapper uses locks to ensure thread-safe access to the COM object, which is critical for concurrent operations.

Event Handling

Query operations use an event-driven architecture:

  • Responses arrive via OnDataResponse callback
  • Events are queued for async processing
  • Tools wait for responses with configurable timeout

Troubleshooting

COM Object Not Found

Error: "Failed to create COM object"

Solution: Ensure CTS Trading API is installed and the COM object is registered. Run the CTS Trading client once to verify installation.

Connection Failures

Error: "Not connected to trading server"

Solution:

  1. Verify appsetting.json has correct server URL
  2. Call initialize() before login()
  3. Call login() before connect()
  4. Check network connectivity

Query Timeout

Note: Query operations may timeout if the server is slow to respond. The default timeout is 5 seconds. This is normal for some operations.

Permission Errors

Ensure you have proper trading permissions and account authorization from CTBC Securities.

Development

Running Automated Tests

The project includes a comprehensive suite of automated tests using pytest and mock.

# Install dev dependencies
uv pip install -e ".[dev]"

# Run all tests
pytest

The tests cover:

  • Models: Validation and serialization of all data structures.
  • Wrapper: Logic for COM object interaction (using mocks).
  • Server: Tool registration and high-level logic.

Manual Testing with Scripts

Logging

The server uses Python's logging module. Set log level via environment variable:

# Windows PowerShell
$env:LOG_LEVEL="DEBUG"
uv run ctbcsec-mcp

# Windows CMD
set LOG_LEVEL=DEBUG
uv run ctbcsec-mcp

Security Considerations

  • Credentials: Never commit credentials to version control
  • Production Use: Use appropriate authentication and authorization
  • Network Security: Ensure secure connection to trading server
  • Access Control: Restrict access to the MCP server appropriately

License

This MCP server is provided for reference and development purposes. Please refer to your CTBC Securities API license agreement for terms of use.

Support

For technical support and API questions, contact CTBC Securities technical support team.

References


Version: 0.1.0
Last Updated: January 21, 2026

推荐服务器

Baidu Map

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。

官方
精选
JavaScript
Playwright MCP Server

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。

官方
精选
TypeScript
Magic Component Platform (MCP)

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。

官方
精选
本地
TypeScript
Audiense Insights MCP Server

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。

官方
精选
本地
TypeScript
VeyraX

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。

官方
精选
本地
graphlit-mcp-server

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。

官方
精选
TypeScript
Kagi MCP Server

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。

官方
精选
Python
e2b-mcp-server

e2b-mcp-server

使用 MCP 通过 e2b 运行代码。

官方
精选
Neon MCP Server

Neon MCP Server

用于与 Neon 管理 API 和数据库交互的 MCP 服务器

官方
精选
Exa MCP Server

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。

官方
精选