Upbit MCP 서버

What is Upbit MCP 서버?

It is an MCP (Model Context Protocol) server implementation for the Upbit cryptocurrency exchange API. It provides tools for market data (ticker, order book, trades, candles), account information, order creation/cancellation, deposit/withdrawal management, technical analysis, backtesting, and chart image generation. It supports SSE (Server-Sent Events) communication for seamless integration with workflow automation tools like n8n.

How to use Upbit MCP 서버?

Deploy using Docker and Docker Compose (provided in the repository). Connect an MCP client to the SSE endpoint exposed by the server. Supply your Upbit API access and secret keys as environment variables or via configuration.

Key features of Upbit MCP 서버

• Real-time market data (ticker, order book, recent trades, candles)
• Account information (balances, open orders, order details, deposit/withdrawal history)
• Order execution and cancellation (limit and market orders)
• Technical analysis indicators (EMA-based MACD, SMA, RSI, Bollinger Bands)
• Backtesting system (SMA crossover, RSI, Bollinger Bands, MACD strategies)
• Chart image generation (candlestick, line, OHLC charts) with web-accessible URLs
• SSE transport for easy n8n and external system integration
• Prompts and resources to guide LLM agent behavior

Use cases of Upbit MCP 서버

• Automate cryptocurrency trading strategies via an LLM agent (e.g., place orders based on technical signals)
• Backtest historical trading strategies (SMA crossover, RSI, Bollinger Bands, MACD) with detailed performance metrics
• Generate chart images for market analysis and share them via web URLs
• Integrate with n8n workflows to monitor market conditions and execute trades
• Retrieve and summarize account portfolio information using natural language prompts

FAQ from Upbit MCP 서버

How does this version differ from the original solangii project?

The original project used stdio transport, making direct integration with n8n difficult. This version adds SSE (Server-Sent Events) support, updates to FastMCP 1.0.0, improves error handling, and enhances the accuracy of technical analysis (using EMA instead of SMA for MACD).

What communication transport does it use?

It uses SSE (Server-Sent Events) over HTTP, designed for integration with external systems like n8n. It does not support stdio transport.

What dependencies are required to run it?

It requires Docker and Docker Compose for deployment, Python with FastMCP 1.0.0, and the Upbit API Python libraries.

Does the backtesting system support natural language requests?

Yes. The backtesting_guide prompt helps convert natural language requests (e.g., "Bitcoin 20-50 SMA crossover on daily for 2024") into the proper tool parameters. The system supports four strategies: SMA crossover, RSI oversold, Bollinger Bands, and MACD signal.

Can I generate chart images from the server?

Yes. The generate_chart_image tool creates candlestick, line, or OHLC charts for a specified market and interval, with optional date ranges