Hyperliquid MCP Server
Enables interaction with the Hyperliquid exchange to trade perpetuals, check positions, and manage risk through natural language. It provides tools for fetching real-time market data, tracking portfolio value, and executing orders with a focus on safety through default paper trading.
README
🚀 Hyperliquid MCP Server
The first MCP server for Hyperliquid — trade perps, check positions, and manage risk via natural language in Claude, ChatGPT, or any MCP-compatible AI agent.
⚡ What You Can Do
Ask Claude in plain English:
"What's my Hyperliquid portfolio worth?" "Show me my open positions and unrealized PnL" "What's the order book for ETH perp right now?" "Which perps have negative funding I can take advantage of?" "How far am I from liquidation on my BTC long?" "Paper trade: buy 0.1 ETH at market" "Set a stop-loss on my SOL position at $120"
🛠️ Tools
| Tool | Description |
|---|---|
get_portfolio_value |
Account balance, equity, margin, unrealized PnL |
get_positions |
All open perp positions with PnL and leverage |
get_orderbook |
Live bid/ask for any market |
get_funding_rates |
Current + predicted funding across all perps |
get_liquidation_price |
Distance from liquidation for each position |
place_order |
Market or limit order (paper trading by default) |
cancel_order |
Cancel one or all open orders |
set_stop_loss |
Set a reduce-only stop-loss trigger |
🚀 Quick Start
1. Clone + Install
git clone https://github.com/YOUR_USERNAME/hyperliquid-mcp
cd hyperliquid-mcp
python -m venv venv
source venv/bin/activate # Windows: venv\Scripts\activate
pip install -r requirements.txt
2. Configure
cp .env.example .env
# Edit .env with your wallet address
# Leave PAPER_TRADING=true until you're ready for live trading
3. Connect to Claude Desktop
Edit ~/Library/Application Support/Claude/claude_desktop_config.json (Mac) or
%APPDATA%\Claude\claude_desktop_config.json (Windows):
{
"mcpServers": {
"hyperliquid": {
"command": "python",
"args": ["/absolute/path/to/hyperliquid-mcp/server.py"],
"env": {
"HL_WALLET_ADDRESS": "0xYOUR_WALLET_ADDRESS",
"HL_PRIVATE_KEY": "0xYOUR_PRIVATE_KEY",
"HL_TESTNET": "true",
"PAPER_TRADING": "true"
}
}
}
}
Restart Claude Desktop. You'll see 🔨 tools available in the input bar.
⚠️ Safety
PAPER_TRADING=true by default. In paper trading mode, all orders are logged but NEVER submitted to Hyperliquid. Safe for testing.
To enable live trading:
- Set
PAPER_TRADING=false - Set
HL_TESTNET=falsefor mainnet - Make sure
HL_PRIVATE_KEYis set - Understand: real money is at risk
The server will always show [PAPER TRADING] or [TESTNET] in tool descriptions so you know which mode you're in.
Never commit your private key. .env is in .gitignore by default.
💰 Monetization (for the managed version)
| Tier | Price | Features |
|---|---|---|
| Open Source | Free | Self-hosted, all tools |
| Cloud Hosted | $29/mo | We host it, no setup, auto-updates |
| Pro Templates | $97 one-time | Strategy templates + backtesting prompts |
| Custom Bot | $500–$2,000 | Done-for-you custom trading agent |
🗺️ Roadmap
- [ ] Historical trade data tool
- [ ] Portfolio P&L over time
- [ ] Multi-account support
- [ ] Telegram bot integration
- [ ] Strategy backtesting tool
- [ ] Automated rebalancing agent
📄 License
MIT — use it, fork it, build on it.
Built by AgentReady — making businesses and markets AI-agent ready.
推荐服务器
Baidu Map
百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。
Playwright MCP Server
一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。
Magic Component Platform (MCP)
一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。
Audiense Insights MCP Server
通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。
VeyraX
一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。
graphlit-mcp-server
模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。
Kagi MCP Server
一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。
e2b-mcp-server
使用 MCP 通过 e2b 运行代码。
Neon MCP Server
用于与 Neon 管理 API 和数据库交互的 MCP 服务器
Exa MCP Server
模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。