How to Trade Crypto with AI on a Regulated Exchange

The LCX Exchange MCP Server connects Claude, ChatGPT, Cursor, Copilot, Gemini, and any MCP-compatible AI client directly to LCX Exchange for natural language crypto trading.

AI is changing how people interact with financial systems. From portfolio tracking to automated strategies, the tools are getting smarter — but the infrastructure behind them often is not. Most AI-powered crypto tools connect to unregulated platforms, lack transparency, or require extensive coding knowledge.

Today, we are making something different available to everyone: the LCX Exchange MCP Server — an open-source tool that connects your AI assistant directly to LCX Exchange for natural language crypto trading.

LCX is the first regulated cryptocurrency exchange to ship a Model Context Protocol (MCP) server. It works with Claude, ChatGPT, Cursor, Windsurf, GitHub Copilot, Gemini, and any MCP-compatible client.

What Is the Model Context Protocol?

The Model Context Protocol (MCP), originally developed by Anthropic, has become the industry standard for connecting AI assistants to external tools and data sources. Think of it as USB-C for AI: one standardized protocol that lets any compatible AI assistant interact with any compatible service.

MCP has been adopted across the industry. Microsoft added MCP support to GitHub Copilot and VS Code. Google integrated it into Gemini CLI. OpenAI brought it to ChatGPT. Cursor and Windsurf support it natively. This broad adoption means that building one MCP server unlocks access from every major AI platform.

How It Works

You talk to your AI assistant in natural language. The assistant calls the MCP server, which translates your intent into authenticated API requests to LCX Exchange. The response flows back as structured data, and your assistant presents it in a clear, readable format.

Your AI (Claude, ChatGPT, Cursor, Copilot, Gemini, )

          |

       stdio (MCP)

          |

    MCP Server (npx lcx-exchange-mcp)

          |

       HTTPS

          |

    LCX Exchange API

What You Can Do: 17 Tools Across Three Categories

Market Data — 7 Tools, No API Keys Required

Anyone can explore LCX markets through AI without creating an account. Just install the MCP server and start asking:

“What trading pairs are available on LCX?”

“What’s the current price of BTC/USDC?”

“Show me the ETH/EUR order book”

“Get 4-hour candles for LCX/EUR for the last week”

Seven tools cover pairs, tickers, order books, public trade history, and OHLCV candlestick data — all live from LCX Exchange.

Tool	Description
lcx_get_pairs	List all trading pairs
lcx_get_pair	Details for a specific pair
lcx_get_tickers	Market overview for all pairs
lcx_get_ticker	Market data for a single pair
lcx_get_order_book	Order book depth (bids and asks)
lcx_get_public_trades	Recent public trade history
lcx_get_kline	OHLCV candlestick data

Trading — 8 Tools, API Keys Required

Once you have your LCX API keys configured, you can trade directly through conversation:

“Buy 500 LCX/EUR at 0.04”

“Sell 0.5 ETH/USDC at market price”

“Change my LCX order to 0.05 EUR”

“Cancel all my open orders”

“Show my trade history for this week”

Eight tools handle order creation (limit and market), modification, cancellation (single or batch up to 25), open order listing, order details, order history, and executed trade history.

Tool	Description
lcx_create_order	Place limit or market order (buy/sell)
lcx_modify_order	Modify an open limit order
lcx_cancel_order	Cancel a single order
lcx_cancel_orders	Cancel up to 25 orders at once
lcx_get_open_orders	List all pending orders
lcx_get_order	Get order details by ID
lcx_get_order_history	Closed and cancelled order history
lcx_get_trade_history	Executed trade history

Account Management — 2 Tools, API Keys Required

Tool	Description
lcx_get_balances	All coin balances
lcx_get_balance	Balance for a specific coin

How to Set Up (Under 2 Minutes)

Prerequisites:

• Node.js 18+ (required to run the MCP server)
• Any supported AI client: Claude Desktop, ChatGPT Desktop, Cursor, Windsurf, VS Code with GitHub Copilot, Gemini CLI, or Claude Code

Step 1: Try Market Data First (No Account Needed)

The fastest way to experience AI trading on LCX. No API keys, no account creation.

Claude Code:

claude mcp add lcx-exchange — npx -y lcx-exchange-mcp

Any other platform: Add the MCP server config without the env block. See the README on GitHub for platform-specific configuration files.

Now open your AI assistant and ask: “What trading pairs are available on LCX?” — you should see live data within seconds.

Step 2: Get Your API Keys (For Trading)

• Go to lcx.com and create an account
• Complete identity verification
• Navigate to Settings → API Management
• Create a new API key
• Save both the API Key and Secret Key (the secret is shown only once)

Step 3: Connect With Trading Access

Claude Code (one command):

claude mcp add lcx-exchange

  -e LCX_API_KEY=your_api_key

  -e LCX_SECRET_KEY=your_secret_key

  — npx -y lcx-exchange-mcp

Claude Desktop — add to your config JSON:

{

  “mcpServers”: {

    “lcx-exchange”: {

      “command”: “npx”,

      “args”: [“-y”, “lcx-exchange-mcp”],

      “env”: {

        “LCX_API_KEY”: “your_api_key”,

        “LCX_SECRET_KEY”: “your_secret_key”

      }

    }

  }

}

ChatGPT, Cursor, Windsurf, GitHub Copilot, Gemini — the same JSON structure works across all platforms. See the full setup guide on GitHub for each.

Build Autonomous Trading Agents

Beyond interactive trading, the MCP server enables fully autonomous agents that trade on your behalf 24/7. Using the Claude Agent SDK or OpenAI Agents SDK, you can build bots that plug directly into the LCX MCP server:

• DCA Bot — Buy a fixed amount of crypto at regular intervals.
• Portfolio Rebalancer — Automatically maintain target asset allocations and rebalance when drift exceeds your threshold.
• Price Alert Bot — Monitor prices and execute trades or send notifications when conditions are met.

Example: a DCA bot that buys 100 LCX (paid in EUR) every 24 hours:

python dca-bot.py –pair LCX/EUR –amount 100 –interval 24h

Complete bot implementations are available in the examples/ directory on GitHub.

Why Regulation Matters for AI Trading

When AI agents execute financial transactions, the compliance status of the underlying platform matters more than ever. An AI assistant placing orders on your behalf should be operating on infrastructure you can trust.

LCX Exchange is licensed by the Financial Market Authority (FMA) of Liechtenstein and holds 8 blockchain-specific regulatory registrations under the Token and Trustworthy Technology Service Provider Act (TVTG). 

This is not a wrapper around an unregulated DEX or a third-party aggregator. Every trade executed through the MCP server goes directly through LCX Exchange’s regulated infrastructure.

Security by design:

• HMAC-SHA256 authentication — every trading request is cryptographically signed.
• API keys stay local — your credentials never leave your machine.
• User-controlled access — revoke API keys at any time from your LCX account.
• Built-in rate limiting — protects against accidental overload (25 req/s market data, 5 req/s trading).

Technical Highlights

For developers who want to understand what is under the hood:

• 17 MCP tools across three categories (market, trading, account)
• Zero external HTTP dependencies — uses Node.js built-in fetch and crypto
• Token-bucket rate limiter — 25 req/s for market data, 5 req/s for trading
• Zod schema validation — every tool input is validated before hitting the API
• Graceful degradation — market tools work without API keys
• TypeScript with strict mode, compiled to ESM

Platform Compatibility

Platform	Status	Config Location
Claude Desktop	Supported	claude_desktop_config.json
Claude Code	Supported	claude mcp add
ChatGPT Desktop	Supported	Settings → Connectors
Cursor	Supported	.cursor/mcp.json
Windsurf	Supported	~/.codeium/windsurf/mcp_config.json
GitHub Copilot (VS Code)	Supported	.vscode/mcp.json
Gemini CLI	Supported	~/.gemini/settings.json
Any MCP Client	Supported	stdio transport

Open Source and Community

The LCX Exchange MCP Server is fully open source under the MIT license. The code is transparent, auditable, and open to contributions. We welcome bug reports, feature requests via GitHub Issues, and pull requests from the community.

A Trading Skill file is also available for Claude users — it teaches the AI best practices for LCX trading, order validation rules, and workflow templates. Copy it to your skills directory and Claude becomes an expert LCX trader out of the box.

What Comes Next

This is version 1.2.0, and community feedback drives what we build next. We are exploring:

• Streaming market data via WebSocket subscriptions
• Advanced portfolio analytics tools
• More autonomous agent templates
• Deeper integration with AI workflow platforms

The vision: AI as a standard interface for regulated crypto trading. No complex UIs, no API boilerplate — just tell your AI what you want.

Get Started Now

Try market data (no account needed):

npx -y lcx-exchange-mcp

Full trading setup in one command (Claude Code):

claude mcp add lcx-exchange

  -e LCX_API_KEY=your_key

  -e LCX_SECRET_KEY=your_secret

  — npx -y lcx-exchange-mcp

Links:

• npm: npmjs.com/package/lcx-exchange-mcp
• GitHub: github.com/tech-lcx/LCX-Exchange-MCP
• LCX Exchange: lcx.com
• MCP Protocol: modelcontextprotocol.io

Start the repo, try it out, and let us know what you build. The future of trading is conversational.