Overview
What is MCP Server for Shioaji?
MCP Server for Shioaji is a Model Context Protocol (MCP) server that gives AI assistants access to the Shioaji trading API for the Taiwanese financial market. It exposes Shioaji functionality as tools, allowing AI models to retrieve stock prices, fetch historical data, and scan stocks by ranking criteria.
How to use MCP Server for Shioaji?
Install with Python 3.10+ and the uv package manager (uv sync). Configure your Shioaji API credentials via environment variables (SHIOAJI_API_KEY, SHIOAJI_SECRET_KEY) or a .env file. Start the server with uv run mcp-server-shioaji; it listens on http://0.0.0.0:8000 by default.
Key features of MCP Server for Shioaji
- Exposes Shioaji tools via the MCP protocol for AI assistants.
- Get current stock prices for one or more symbols.
- Fetch K‑bar (candlestick) historical data with configurable date ranges.
- Scan and rank stocks by volume, amount, tick count, price change, and more.
- Supports ascending/descending sort and adjustable result limits.
Use cases of MCP Server for Shioaji
- An AI assistant queries the latest price of a portfolio of Taiwanese stocks.
- A trading bot fetches historical candlestick data for backtesting.
- A market screener ranks stocks by trading volume to identify active stocks.
FAQ from MCP Server for Shioaji
What dependencies does the server require?
Python 3.10 or higher and the uv package manager. Shioaji API credentials are also required.
How do I authenticate with the Shioaji API?
Set the SHIOAJI_API_KEY and SHIOAJI_SECRET_KEY environment variables or provide them in a .env file in the server root directory.
Where is the financial data stored?
The server does not store data locally; it fetches live data from the Shioaji API on each request.
What transport does the server use?
The server runs as an HTTP server (default port 8000) and communicates via the Model Context Protocol (MCP).
Are there any known limits on the number of requests or symbols?
The README does not specify rate limits, but it relies on the Shioaji API’s own constraints. The scan_stocks tool defaults to a limit of 20 results.
