TigerOpen Python SDK

Before using the SDK, you must register as a developer at the Developer Information Page to obtain your tiger_id, private_key (RSA PKCS#1 format), and account number.

There are two ways to configure the client:

Method 1: Configuration File

Create a file named tiger_openapi_config.properties (e.g., in ~/.tigeropen/) with the following content:

tiger_id=your_tiger_id
private_key_pk1=your_pkcs1_private_key
account=your_account

Then initialize the config in Python:

from tigeropen.tiger_open_config import TigerOpenClientConfig
config = TigerOpenClientConfig(props_path='~/.tigeropen/')

Method 2: Direct Code Configuration

Pass credentials directly to the TigerOpenClientConfig object:

from tigeropen.tiger_open_config import TigerOpenClientConfig

config = TigerOpenClientConfig()
config.tiger_id = 'your_tiger_id'
config.account = 'your_account'
config.private_key = 'your_private_key_content'

The Model Context Protocol (MCP) Server allows AI tools like Cursor, Claude Code, and Trae to perform market queries and trade management via natural language.

Installation

First, ensure uv is installed:

pip install uv

Configuration

Add the following to your AI tool's MCP settings:

For Personal Accounts:

{
  "mcpServers": {
    "tigermcp": {
      "command": "uvx",
      "args": ["tigermcp"],
      "env": {
        "TIGEROPEN_TIGER_ID": "your_tiger_id",
        "TIGEROPEN_PRIVATE_KEY": "your_private_key",
        "TIGEROPEN_ACCOUNT": "your_account"
      }
    }
  }
}

For Institutional Accounts:

{
  "mcpServers": {
    "tigermcp": {
      "command": "uvx",
      "args": ["tigermcp"],
      "env": {
        "TIGEROPEN_TIGER_ID": "your_tiger_id",
        "TIGEROPEN_PRIVATE_KEY": "your_private_key",
        "TIGEROPEN_ACCOUNT": "your_account",
        "TIGEROPEN_SECRET_KEY": "your_secret_key",
        "TIGEROPEN_TOKEN": "your_token"
      }
    }
  }
}

Note:

• To enable Read-only mode (disables order placement), add "TIGERMCP_READONLY": "true" to the env block.
• macOS 12 users may need to install coreutils (brew install coreutils) if encountering realpath errors.

The Tiger MCP Server requires TigerOpen API credentials to function. You can provide these in two ways:

Option 1: Using a properties file (Recommended)

Set the TIGEROPEN_PROPS_PATH environment variable to the absolute path of your tiger_openapi_config.properties file (which should contain your tiger_id, private_key, and account information).

Option 2: Using direct environment variables

Set the following variables directly:

• TIGEROPEN_TIGER_ID: Your Tiger ID
• TIGEROPEN_PRIVATE_KEY: Your private key
• TIGEROPEN_ACCOUNT: Your trading account

Read-Only Mode

To prevent the server from performing any trading actions, set the TIGERMCP_READONLY environment variable to true.

export TIGEROPEN_PROPS_PATH="path/to/your/tiger_openapi_config.properties"
# OR
export TIGEROPEN_TIGER_ID="your Tiger ID"
export TIGEROPEN_PRIVATE_KEY="your private key"
export TIGEROPEN_ACCOUNT="your trading account"
# For read-only mode
export TIGERMCP_READONLY="true"

TigerOpen provides a Model Context Protocol (MCP) Server that allows AI coding tools like Cursor, Claude Code, or Trae to query market data and manage trades using natural language.

Installation

Ensure uv is installed, then use uvx to run the server.

pip install uv

Configuration for AI Tools

Add the following to your tool's MCP configuration file:

For Personal Accounts:

{
  "mcpServers": {
    "tigermcp": {
      "command": "uvx",
      "args": ["tigermcp"],
      "env": {
        "TIGEROPEN_TIGER_ID": "your_tiger_id",
        "TIGEROPEN_PRIVATE_KEY": "your_private_key",
        "TIGEROPEN_ACCOUNT": "your_account"
      }
    }
  }
}

For Institutional Accounts:

{
  "mcpServers": {
    "tigermcp": {
      "command": "uvx",
      "args": ["tigermcp"],
      "env": {
        "TIGEROPEN_TIGER_ID": "your_tiger_id",
        "TIGEROPEN_PRIVATE_KEY": "your_private_key",
        "TIGEROPEN_ACCOUNT": "your_account",
        "TIGEROPEN_SECRET_KEY": "your_secret_key",
        "TIGEROPEN_TOKEN": "your_token"
      }
    }
  }
}

Tips

• Read-Only Mode: To prevent the AI from placing trades, add "TIGERMCP_READONLY": "true" to the env block.
• macOS Users: If you encounter realpath errors, install coreutils via brew install coreutils.