hpsilab-mcp 0.1.0

Python SDK for H|ψ Quantum Finance APIs.

H|ψ⟩ Quantum Finance MCP SDK

AI-powered quantitative finance tools via MCP.

Connect Claude, Codex, and AI agents directly to:

https://hpsilab.com/mcp

Features:

• AI Predictions
• IV Radar
• Option Pressure
• Monte Carlo Simulation
• Equity Curves

Project Overview

H|ψ⟩ Quantum Finance MCP provides AI-powered quantitative finance tools through a hosted MCP endpoint. Client applications can use MCP to request structured stock analysis, AI prediction summaries, implied-volatility context, option pressure levels, Monte Carlo simulation ranges, equity curve metrics, and report assets.

This repository contains open-source SDK skeletons and examples for connecting MCP-compatible clients to the hosted H|ψ⟩ Quantum Finance MCP service.

This repository includes:

• Python SDK skeleton in python/client.py
• TypeScript SDK skeleton in typescript/client.ts
• Client setup examples in examples/
• Documentation for available MCP tools and prompt patterns

Hosted MCP Endpoint

H|ψ⟩ Quantum Finance provides a hosted MCP endpoint:

https://hpsilab.com/mcp

No self-hosting required.

Connect directly from Claude Desktop, Codex, or other MCP-compatible clients.

Quick Example

Once connected, ask:

• Analyze NVDA
• Predict IONQ
• Show IV Radar for QBTS
• Compare SOUN vs PLTR

Setup

Configure your MCP-compatible client to connect to the hosted endpoint:

https://hpsilab.com/mcp

Typical MCP client configuration requires:

• The hosted MCP endpoint URL
• Any required client-side authentication configured outside this repository
• A client that supports MCP tool calls

Do not place secrets directly in this repository. Use your MCP client's supported secret or environment configuration mechanism.

Quick Start

Python

from client import HpsiMcpClient

client = HpsiMcpClient(server_url="https://hpsilab.com/mcp")

analysis = client.call_tool(
    "analyze_stock",
    {"symbol": "NVDA"},
)

print(analysis)

TypeScript

import { HpsiMcpClient } from "./client";

const client = new HpsiMcpClient({
  serverUrl: "https://hpsilab.com/mcp",
});

const analysis = await client.callTool("analyze_stock", {
  symbol: "NVDA",
});

console.log(analysis);

The SDK files are skeletons. Adapt transport, authentication, retries, and response validation to your runtime and MCP client stack.

MCP Setup

An MCP client registers the hosted H|ψ⟩ endpoint under a named server entry, then exposes its tools to the assistant.

Example endpoint:

https://hpsilab.com/mcp

Use your MCP client's supported remote-server configuration. Avoid committing credentials or local secrets.

Claude Desktop Setup

Add the hosted endpoint to your Claude Desktop MCP configuration file.

Example:

{
  "mcpServers": {
    "hpsilab": {
      "command": "npx",
      "args": [
        "-y",
        "mcp-remote",
        "https://hpsilab.com/mcp"
      ]
    }
  }
}

Restart Claude Desktop after changing the MCP configuration. Once connected, Claude should be able to discover the H|ψ⟩ Quantum Finance MCP tools.

See examples/claude-desktop.md for a fuller setup template.

Available Tools

Tool names may vary by deployment, but the public MCP surface is expected to include:

• analyze_stock: Aggregate stock research signal and summary data.
• generate_stock_images: Generate stock-report chart images.
• generate_stock_research_report: Generate a markdown research report with chart embeds.
• get_ai_prediction: Retrieve an AI-powered next-day directional prediction summary.
• get_equity_curves: Retrieve equity curve and backtest performance metrics.
• get_iv_radar: Retrieve implied-volatility structure and option sentiment context.
• get_monte_carlo: Retrieve a short-horizon Monte Carlo simulation summary.
• get_option_pressure: Retrieve option-chain pressure levels such as max pain and gamma context.

These tools return research-oriented information. They are not financial advice.

Example Prompts

• "Analyze NVDA using the H|ψ⟩ Quantum Finance MCP tools and summarize the bullish and bearish factors."
• "Generate a stock research report for AAPL with charts."
• "Check IV Radar and Option Pressure for SPY."
• "Compare the Monte Carlo Simulation outlook and AI Prediction for TSLA."
• "Show the recent Equity Curves metrics for my watchlist."

Repository Scope

This repository is for open-source client examples only. It deliberately excludes:

• Proprietary algorithms
• Internal model weights or prompts
• Trading strategy implementation details
• Server-side business logic
• Secrets, credentials, or private deployment configuration

License

MIT. See LICENSE.