π Luno MCP Server - FastMCP 2.0
@amanasmuei
mcp server for luno
Overview
What is π Luno MCP Server - FastMCP 2.0?
A modern Model Context Protocol (MCP) server for the Luno cryptocurrency exchange, built with FastMCP 2.0 and Python 3.12. It provides tools for real-time prices, historical candlestick data, account management, and trading. Designed for Claude Desktop users.
How to use π Luno MCP Server - FastMCP 2.0?
Configure Claude Desktop by adding the server to claude_desktop_config.json with the Python virtual environment path, server script, and Luno API credentials as environment variables (LUNO_API_KEY, LUNO_API_SECRET). Restart Claude Desktop and ask natural language queries such as βWhat's the Bitcoin price in EUR?β or βShow me my account balanceβ.
Key features of π Luno MCP Server - FastMCP 2.0
- Real-time cryptocurrency prices for any trading pair
- Historical OHLC candlestick data with multiple timeframes
- Price range analysis over 1β30 days
- Account balance and transaction history tools
- Place, cancel, and check order status
- Trading fees query tool
Use cases of π Luno MCP Server - FastMCP 2.0
- Get realβtime BTC/EUR price via natural language
- Retrieve historical XBTZAR candlestick data for analysis
- Perform price range analysis for the past 7 or 30 days
- Check account balances and recent transactions
- Place or cancel buy/sell orders through Claude
FAQ from π Luno MCP Server - FastMCP 2.0
What dependencies are required?
Python 3.12, FastMCP 2.5.1, httpx, pydantic, and python-dotenv. All are installed in the provided virtual environment.
How are API credentials handled?
Credentials are stored as environment variables (LUNO_API_KEY and LUNO_API_SECRET) in the Claude Desktop config. All communications use HTTPS, and no credentials are logged or exposed.
Which tools require API credentials?
Public tools (get_crypto_price, get_market_overview) need no credentials. Historical data tools and all private tools require valid Luno API credentials.
What transport protocol is used?
STDIO transport with JSONβRPC 2.0.
Are there any request limits?
Historical candlestick data is limited to 1000 candles per request. Price range analysis supports periods of 1β30 days.