Alpaca MCP Server
Trading API access for stocks, options, and crypto directly in your AI environment.
Data last scanned today · Reviewed today
Overview
The Alpaca MCP Server provides a direct interface for interacting with Alpaca's trading platform through AI assistants. Built on FastMCP and OpenAPI, this tool allows for natural language management of portfolios and access to real-time market data across various asset classes. Version 2 represents a complete structural rewrite of the previous iteration. Users should be aware that this update introduces significant changes to tool names, parameters, and schema definitions compared to the original version.
Our verdict
With a security score of 66/100, this server is categorized as medium risk. While it benefits from being actively maintained by the official Alpaca organization, users should be aware that the server currently lacks a read-only mode, meaning all installed configurations grant full write access to the API. It is best suited for experienced traders who are comfortable managing API-level permissions manually.
- Maintained directly by the official Alpaca organization.
- Active development with commits as recent as one day ago.
- Zero direct dependencies minimize the supply-chain attack surface.
- Sixteen individual contributors help reduce bus-factor risk.
- Full write access is granted to all installs.
- Lacks support for OAuth-based authentication.
- Not yet undergone a formal editorial security review.
Tools
| Tool | Description | Risk |
|---|---|---|
| get_account_info | Balance, margin, and account status | low |
| get_account_config | Trading restrictions, margin settings, PDT checks | low |
| update_account_config | Update account configuration settings | medium |
| get_portfolio_history | Equity and P/L over time | low |
| get_account_activities | Fills, dividends, transfers | low |
| get_account_activities_by_type | Activities filtered by type | low |
| get_orders | Retrieve orders with filters | low |
| get_order_by_id | Single order by ID | low |
| get_order_by_client_id | Single order by client order ID | low |
| replace_order_by_id | Replace an existing open order | medium |
Compatibility
| Client | Local | Docker | Remote | Read-only |
|---|---|---|---|---|
| Claude Desktop | ||||
| Cursor | ||||
| VS Code | ||||
| Windsurf | ||||
| ChatGPT |
Frequently asked questions
›Can I upgrade directly from V1 to V2?
No, V2 is a complete rewrite and is not a drop-in replacement. Tool names, parameters, and configuration methods have changed, so you must treat it as a new installation.
›What are the core technical requirements for running the server?
You need Python 3.10 or higher and the uv package manager installed. You also require active Alpaca Trading API keys, which work with a free paper trading account.
›How can I limit the tools available to my AI assistant?
You can restrict capabilities by using the ALPACA_TOOLSETS environment variable to define which specific toolsets the server should expose.
›What should I do if my MCP client shows old V1 tools after upgrading?
You must restart your MCP client (such as Claude Desktop or Cursor) and start a fresh chat session to ensure the client clears its cache and fetches the new V2 tool definitions.
›Can I continue using V1 if my workflow relies on it?
Yes, you can stay on V1 by pinning your installation to the last V1 release version (e.g., uvx alpaca-mcp-server==1.x.x serve) in your MCP client configuration.
Badge
Maintain this server? Add the live badge to your README.