horus-flow-mcp
🛑 停止预测 K 线。开始测量引力。
散户交易者使用滞后指标(RSI/MACD)根据“过去”来猜测市场走势。Horus 利用 Level 2 订单簿物理特性、Tick 不平衡和 5 秒流增量,精确测量机构巨鲸“当下”的动作。
Horus 不会问“我们是否超买?”。它测量引力并告诉你:“巨鲸正在伪造买单,激进的吃单者正在击穿卖单。流动性正在崩溃。立即撤离。”
⚡ 为什么 AI 智能体(Claude, Cursor)钟爱 Horus
如果你将 AI 智能体连接到基础技术指标源,它会被噪音干扰。 如果你为 AI 智能体提供 Horus,它将获得机构级的决策矩阵。
看看 Horus 在一次 流动性撤出 / 闪崩 尝试中实时捕捉到的内容:
{
"symbol": "BTCUSDT",
"signal": "LIQUIDITY_EVENT",
"confidence": 0.99,
"market_state": "DISTRIBUTION",
"risk": "EXTREME",
"description": "Liquidity withdrawn under price. Aggressive taking.",
"metrics": {
"bid_ask_ratio": 7.934,
"buy_sell_ratio": 0.393,
"delta_5s": -83040.05,
"whale_activity": true,
"large_sell_count": 4,
"delta_accel": 3.8,
"wall_side": "BID",
"wiseman_climate": {
"market_mode": "CHOP",
"health": "FRAGILE",
"confidence": 0.85
},
"flags": [
"GLOBAL_LIQUIDITY_EVENT",
"SPOOFING_DETECTED(wall=BID)"
]
},
"timestamp": 1776107738.975
}机构级 Alpha:
bid_ask_ratio: 7.934&wall_side: BID:巨鲸在市场下方放置了巨大的伪造买单墙,以制造虚假支撑。buy_sell_ratio: 0.393:Horus 看穿了伪造行为(SPOOFING_DETECTED标志)。它测量真实的“吃单”流,并发现激进的卖出正在摧毁订单簿。delta_accel: 3.8&whale_activity: true:卖出动能加速了 3.8 倍,原生追踪到毫秒级内的 4 次重大巨鲸抛售。wiseman_climate: FRAGILE:与 Horus SaaS 宏观大脑完美集成,在攻击前验证比特币的整体环境是否脆弱。结论:凭借
0.99的置信度,AI 引擎触发LIQUIDITY_EVENT。它抢在随后的 1 分钟崩盘前采取行动。这是机构级的表现。
🏗️ 架构
graph TD
%% Styling
classDef crypto_stream fill:#F3BA2F,stroke:#000,color:#000,stroke-width:2px;
classDef equity_stream fill:#000,stroke:#09b533,color:#09b533,stroke-width:2px;
classDef compute fill:#1A1F36,stroke:#00D6FF,color:#fff,stroke-width:2px;
classDef mcp fill:#632CA6,stroke:#fff,color:#fff,stroke-width:2px;
classDef client fill:#FF3366,stroke:#fff,color:#fff,stroke-width:2px;
%% Ingestion
subgraph Data_Pipelines [Sub-Millisecond Websocket Ingestion]
B[Binance WSS <br/> L1/L2 Book]:::crypto_stream
A[Alpaca WSS <br/> SIP Equities]:::equity_stream
end
%% Engine
subgraph Core_Engine [The Physics Engine]
IC[Imbalance Calculator <br/> Bid/Ask Spread]:::compute
FC[Flow Calculator <br/> Tape Deltas]:::compute
Tkr[Prediction Tracker <br/> In-Memory Evaluator]:::compute
BC[Behavioral Court <br/> Spoofing & Liquidity Rules]:::compute
end
%% Output
subgraph Output_Layer [Data Shield & Delivery]
MCP[AI Agent MCP Context <br/> Context-Aware Prompts]:::mcp
API[FastAPI Client <br/> Safe Shielded Outputs]:::client
Dash[Real-time Dashboard <br/> Live Edge Proof]:::client
end
B --> IC
A --> IC
B --> FC
A --> FC
IC --> BC
FC --> BC
BC --> Tkr
Tkr -- "Validates 1M accuracy" --> BC
BC --> MCP
BC --> API
BC --> Dash🚀 交易机器人快速入门
开始使用 Horus 只需不到 60 秒。 ⚡ 立即测试: 下载我们的 Postman Collection 以直接从浏览器 ping API。
# 1. Install Dependencies
pip install -r requirements.txt
# 2. Fire up the Core Engine
uvicorn app.main:app --host 0.0.0.0 --port 8011在你的本地机器人/智能体中,只需拉取任何加密货币或美股的实时流:
import requests
response = requests.get(
"http://127.0.0.1:8011/v1/flow/crypto/BTCUSDT",
headers={"X-API-Key": "your_secure_api_key"}
)
flow_data = response.json()
# The Zero-Human Decision Loop:
if flow_data['signal'] in ['EMERGENCY_DUMP', 'BAILOUT', 'LIQUIDITY_EVENT']:
print("Market gravity collapsing. Market selling all positions.")
# execute_market_sell()🌐 实时仪表盘
Horus 开箱即用,配备了一个令人惊叹的玻璃拟态 WebSocket 仪表盘 (/dash/)。
它具有一个 本地边缘验证追踪器,严格测量由物理引擎生成的 1 分钟 K 线方向预测,并公开透明地展示胜率。
为那些需要看清矩阵的人打造,追求绝对精度。 — HORUS INTELLIGENCE 👁️
Latest Blog Posts
MCP directory API
We provide all the information about MCP servers via our MCP API.
curl -X GET 'https://glama.ai/api/mcp/v1/servers/horustechltd/horus-flow-mcp'
If you have feedback or need assistance with the MCP directory API, please join our Discord server