Alpaca's Official MCP Server

What is Alpaca's Official MCP Server?

A Model Context Protocol (MCP) server for Alpaca's Trading API that enables natural language trading operations through AI assistants like Claude, Cursor, and VS Code. It supports stocks, options, crypto, portfolio management, and real-time market data.

How to use Alpaca's Official MCP Server?

Install prerequisites (Python 3.10+, uv, Alpaca API keys, an MCP client) and run uvx alpaca-mcp-server init or python3 install.py. Then configure your MCP client by adding the server to its JSON config file with your API keys in the env section.

Key features of Alpaca's Official MCP Server

• Real-time and historical market data for stocks, crypto, and options
• Account balance, buying power, and position inspection
• Place and cancel market, limit, stop, and trailing-stop orders
• Single-leg and multi-leg options trading (spreads, straddles)
• Crypto market, limit, and stop-limit order support
• Market status and corporate actions data

Use cases of Alpaca's Official MCP Server

• Ask an AI assistant to check your portfolio's current balance and positions
• Place a market order for stocks or crypto using natural language
• Search and trade options contracts by expiration, strike, and type
• Liquidate all or a percentage of a position through a text command

FAQ from Alpaca's Official MCP Server

What are the runtime requirements?

Python 3.10 or higher, the uv package manager, and a compatible MCP client (Claude Desktop, Cursor, VS Code, etc.).

How do I get API keys and switch between paper and live trading?

Generate free paper trading API keys from the Alpaca dashboard. To switch to live trading, update the ALPACA_API_KEY, ALPACA_SECRET_KEY, and set ALPACA_PAPER_TRADE=False in both the .env file and your MCP client config JSON.

What financial instruments are supported?

Stocks, ETFs, crypto, and options (single-leg and multi-leg strategies) across US markets.

Where do my API credentials and data live?

All credentials are stored locally in your .env file and MCP client config JSON. The server can run locally or be deployed remotely (e.g., with Docker or to a cloud service for Claude Mobile/ChatGPT).

Does the server support OAuth or remote HTTP transport?

Yes, the README documents both OAuth Bearer Token Support and HTTP Transport for remote usage.