Liquidiction
Query live HIP-4 prediction market data on Hyperliquid - real-time odds, orderbooks, trades, candles, user positions, and portfolio analysis via 10 MCP tools.
README
<p align="center"> <img src="https://testnet.liquidiction.xyz/logo.png" alt="Liquidiction" width="80" /> </p>
<h1 align="center">Liquidiction MCP Server</h1>
<p align="center"> Query live Hyperliquid HIP-4 prediction market data from any MCP-compatible AI agent </p>
<p align="center"> <a href="https://testnet.liquidiction.xyz">App</a> | <a href="https://liquidiction.xyz">Website</a> | <a href="https://x.com/LiquidictionHL">Twitter</a> </p>
What is this?
An MCP (Model Context Protocol) server that gives AI agents like Claude, GPT, and custom agents direct access to live HIP-4 prediction market data on Hyperliquid. This is the first MCP server for prediction markets.
No API keys required. Connects directly to Hyperliquid's public API.
Tools (10)
| Tool | Description |
|---|---|
list_markets |
All active prediction markets with current prices |
get_market_detail |
Detailed info for a specific outcome ID |
get_market_summary |
Rich overview with probabilities, settlement status, and parsed metadata |
get_orderbook |
L2 order book (bids, asks, spread) for an outcome side |
get_prices |
Current mid-prices for all outcome coins |
get_candles |
OHLCV candle data for charting and analysis |
get_recent_trades |
Latest trades on a specific market |
get_user_fills |
Trade history for a wallet address |
get_open_orders |
Resting limit orders for a wallet |
get_user_positions |
Current outcome share holdings and values |
Quick Start
Claude Desktop
Add to your Claude Desktop config (claude_desktop_config.json):
{
"mcpServers": {
"liquidiction": {
"command": "npx",
"args": ["tsx", "/path/to/liquidiction-mcp/mcp-server.ts"]
}
}
}
Restart Claude Desktop, then try:
- "What are the current HIP-4 prediction market odds?"
- "Show me the orderbook for outcome 13"
- "What positions does 0x1393...1869 hold?"
Claude Code
claude mcp add liquidiction npx tsx /path/to/liquidiction-mcp/mcp-server.ts
Run Standalone
git clone https://github.com/Liquidiction/liquidiction-mcp.git
cd liquidiction-mcp
npm install
npm start
Configuration
| Env Variable | Default | Description |
|---|---|---|
HL_API_URL |
https://api.hyperliquid-testnet.xyz |
Hyperliquid API endpoint |
Set to https://api.hyperliquid.xyz for mainnet data.
Example Queries
Market Odds
"What are the odds on BTC above $71,566?"
Uses get_market_summary to fetch all markets, finds BTC markets, returns current probabilities.
Orderbook Analysis
"Is there good liquidity on the Hypurr vs Usain Bolt market?"
Uses get_orderbook to check bid/ask depth, spread, and order count.
Portfolio Check
"What are my positions and PnL?"
Uses get_user_positions + get_user_fills to show holdings and trade history.
Price History
"How has the HYPE prediction market moved in the last 24 hours?"
Uses get_candles with 1h interval to show price movement over time.
What is HIP-4?
HIP-4 (Hyperliquid Improvement Proposal 4) introduces prediction markets that run natively on Hyperliquid's L1 order book engine. Key features:
- Binary outcomes (Yes/No) priced 0-100 cents
- Fully collateralized (no leverage, no liquidation risk)
- Shares the same infrastructure as perps and spot
- Supports recurring markets (15M, 1H, 1D periods)
- Settlement via on-chain oracle
Premium API (x402)
For enriched data with orderbook depth, volume, and metadata:
GET https://testnet.liquidiction.xyz/api/premium/analytics
Paywalled at $0.001/request via x402 USDC micropayments on Base Sepolia.
Links
- App: testnet.liquidiction.xyz
- Landing: liquidiction.xyz
- Twitter: @LiquidictionHL
- Telegram: t.me/Liquidiction
- Built on: Hyperliquid
License
MIT
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。