MCP Server

Use the Tiger OpenAPI MCP Server on any platform that supports the Model Context Protocol (MCP). The server is built on the Tiger OpenAPI SDK.
If you have not activated Tiger OpenAPI, follow Preparation to apply for access and obtain your Tiger ID, private key, and other required credentials.

Market data access: Real-time market data, including quotes, ticks, and depth, requires a separate Tiger OpenAPI market data subscription and is independent of access in the Tiger Trade app. Historical data, such as candlestick bars (K-line data), is available according to your subscription tier. See Market Data Permissions for details.

Install uv

On macOS or Linux, open a terminal and run:

curl -LsSf https://astral.sh/uv/install.sh | sh

uv is a Python package and environment manager used here to run the MCP Server.

Add the MCP Server

Open the MCP settings in your client, select Manual Configuration, and add the following configuration.

Replace the Tiger ID, private key, and account ID placeholders with your actual values. To run the server in read-only mode (with trading operations disabled), set TIGERMCP_READONLY to true; otherwise, set it to false. Alternatively, use TIGEROPEN_PROPS_PATH to point to your properties file instead of providing the credential values separately.

{
  "mcpServers": {
    "tigermcp": {
      "command": "uvx",
      "args": [
        "--python",
        "3.13",
        "tigermcp"
      ],
      "env": {
        "TIGEROPEN_TIGER_ID": "your tiger id",
        "TIGEROPEN_PRIVATE_KEY": "your private key",
        "TIGEROPEN_ACCOUNT": "your account id",
        "TIGERMCP_READONLY": "true"
      }
    }
  }
}

Replace the Tiger ID, private key, secret key, and account ID placeholders with your actual values. Hong Kong clients must also replace the TBHK 2FA token placeholder; clients in other regions must remove that line. To run the server in read-only mode with trading operations disabled, set TIGERMCP_READONLY to true; otherwise, set it to false. Alternatively, use TIGEROPEN_PROPS_PATH to point to your properties file instead of providing the credential values separately.

{
  "mcpServers": {
    "tigermcp": {
      "command": "uvx",
      "args": [
        "--python",
        "3.13",
        "tigermcp"
      ],
      "env": {
        "TIGEROPEN_TIGER_ID": "your tiger id",
        "TIGEROPEN_PRIVATE_KEY": "your private key",
        "TIGEROPEN_SECRET_KEY": "your secret key",
        "TIGEROPEN_ACCOUNT": "your account id",
        "TIGEROPEN_TOKEN": "TBHK 2fa token",
        "TIGERMCP_READONLY": "true"
      }
    }
  }
}

Troubleshoot MCP Server Errors on macOS 12 or Earlier

On macOS 12 or earlier, you may encounter the following error when running the MCP Server:

/Users/tiger/.cache/uv/archive-v0/5iV7KVbKUQlypQW-eBHBn/bin/tigermcp: line 2: realpath: command not found
/Users/tiger/.cache/uv/archive-v0/5iV7KVbKUQlypQW-eBHBn/bin/tigermcp: line 2: /Users/tiger/python: No such file or directory
/Users/tiger/.cache/uv/archive-v0/5iV7KVbKUQlypQW-eBHBn/bin/tigermcp: line 2: exec: /Users/tiger/python: cannot execute: No such file or directory

This error occurs because the realpath command is unavailable.
First, open a terminal and run brew -v to check whether Homebrew is installed. If the command is unavailable, install Homebrew:

/bin/bash -c "$(curl -fsSL https://raw.githubusercontent.com/Homebrew/install/HEAD/install.sh)"

Then use Homebrew to install coreutils:

brew install coreutils

After installation, run the MCP Server again.

Disclaimer

TIGER MCP is solely a tool for connecting AI applications to the TIGER API. Its output depends on the behavior of the AI tools and the capabilities of the connected large language models (LLMs). Differences in reasoning, classification, and data-processing capabilities, as well as platform-specific limitations, may affect the accuracy and completeness of results. Risks include, but are not limited to, misinterpreted times, data-analysis errors, truncated or omitted information, and tool invocation failures.
API users must independently assess all investment decisions and other actions taken through TIGER MCP. Users are expected to have the necessary knowledge and to implement appropriate security and access controls. Each TIGER API user is solely responsible for all associated risks and liabilities. We make no warranties regarding the accuracy, completeness, reliability, or timeliness of output from AI tools or LLMs, and we accept no responsibility for any resulting legal liability, financial loss, or tax consequence.