Skip to main content

Python SDK for the HPSILab quantitative finance API — options analytics, IV surface modeling, Monte Carlo simulation, and AI-driven predictions, with native MCP support.

Project description

HPSILab Python REST SDK

hpsilab-mcp is the official Python SDK for the hosted hpsilab.com REST API — quantitative finance and options analytics (IV surface, Monte Carlo simulation, AI predictions, pre-trade risk scans, and more).

Note: This package wraps REST endpoints only. It does not implement MCP transport — see MCP Transport below if you need that.

Requirements

  • Python >= 3.9

Installation

pip install hpsilab-mcp

To let the client pay per call when the free quota runs out (see Paying past the free quota):

pip install "hpsilab-mcp[x402]"

Quick Start

from hpsilab_mcp import HpsiMcpClient

client = HpsiMcpClient()

calls = {
    "analyze_stock": client.analyze_stock("NVDA"),
    "get_ai_prediction": client.get_ai_prediction("NVDA"),
    "get_iv_radar": client.get_iv_radar("NVDA"),
    "get_option_pressure": client.get_option_pressure("NVDA"),
    "get_pretrade_risk_scan": client.get_pretrade_risk_scan("NVDA"),
    "get_monte_carlo": client.get_monte_carlo("NVDA"),
    "get_equity_curve": client.get_equity_curve("NVDA"),
    "generate_stock_images": client.generate_stock_images("NVDA"),
    "generate_stock_research_report": client.generate_stock_research_report("NVDA"),
}

print(calls["analyze_stock"])

Authentication

All listed REST SDK methods are callable without an API key. The SDK does not block any method client-side; any access restrictions are enforced server-side.

from hpsilab_mcp import HpsiMcpClient

client = HpsiMcpClient(
    api_key="YOUR_API_KEY",
    base_url="https://hpsilab.com",
)

result = client.get_ai_prediction("TSLA")
print(result)

Pass an api_key to raise rate limits or unlock account-specific features, where applicable.

Paying past the free quota

Anonymous callers get a per-tool daily quota. Once it runs out — or when calling a Pro tool, which has no anonymous allowance — the API answers HTTP 402 with an x402 payment challenge instead of refusing outright, so an agent can pay for the call in USDC on Base and keep working.

Without a wallet the SDK raises HpsiMcpPaymentError with the challenge attached, and you can pay it however you like:

from hpsilab_mcp import HpsiMcpClient, HpsiMcpPaymentError

try:
    client.get_monte_carlo("NVDA")
except HpsiMcpPaymentError as exc:
    print(exc.tool, exc.price)   # get_monte_carlo $0.10
    print(exc.accepts)           # scheme / network / asset / amount / payTo

With a wallet, the client signs the challenge and repeats the request for you:

from hpsilab_mcp import HpsiMcpClient, X402Wallet

client = HpsiMcpClient(wallet=X402Wallet(PRIVATE_KEY, max_price_usdc=0.20))
client.get_monte_carlo("NVDA")   # free while quota lasts, paid after that

X402Wallet() reads HPSILAB_X402_PRIVATE_KEY when no key is passed, and a client with no wallet= picks that variable up automatically. Three things hold regardless of how it is configured:

  • Nothing is paid pre-emptively — the free call is always attempted first.
  • Each call is retried once after paying; a server that keeps answering 402 cannot drain the wallet.
  • max_price_usdc (default $1.00) caps a single call. A challenge asking for more is refused, not signed.

Signing happens locally; the private key never leaves your process.

Tool Price per call
analyze_stock, get_pretrade_risk_scan $0.15
get_monte_carlo $0.10
get_iv_radar, get_option_pressure, get_ai_prediction, get_equity_curve $0.05
generate_stock_research_report $0.35

Version

import hpsilab_mcp

print(hpsilab_mcp.__version__)

Request tracking

Every request carries X-HPSILAB-Source: sdk, X-HPSILAB-Client: python-sdk, X-HPSILAB-Version: <installed version>, X-HPSILAB-Tool: <method name> (e.g. get_ai_prediction), and User-Agent: hpsilab-python-sdk/<version>. These are merged into any custom headers you pass to HpsiMcpClient(...) without overriding Authorization or other business headers.

REST SDK Methods

Method Endpoint
analyze_stock(symbol) GET /api/analyze_stock/{symbol}
get_ai_prediction(symbol) GET /api/ai_prediction/{symbol}
get_iv_radar(symbol) GET /api/iv_batch?symbols={symbol}
get_option_pressure(symbol) GET /api/option_pressure/{symbol}
get_pretrade_risk_scan(symbol) GET /api/pretrade-risk-scan?symbol={symbol}
get_monte_carlo(symbol) GET /api/monte_carlo/{symbol}
get_equity_curve(symbol) GET /api/equity_curve/{symbol}
get_equity_curves(symbol) Deprecated alias of get_equity_curve — warns on use
generate_stock_images(symbol) POST /api/stock_report/{symbol}/images
generate_stock_research_report(symbol) POST /api/stock_report/{symbol}/research_report

The SDK itself does not publish MCP tool annotations; those are declared by the MCP server in its tools/list response. The GET analysis methods are read-only. The two generate_* methods create or refresh hosted artifacts and may consume quota or trigger payment; repeated calls are not guaranteed to be idempotent.

Capability Matrix

Capability REST API MCP
analyze_stock Yes Yes
get_ai_prediction Yes Yes
get_iv_radar Yes Yes
get_option_pressure Yes Yes
get_pretrade_risk_scan Yes Yes
get_monte_carlo Yes Yes
get_equity_curve Yes Yes
generate_stock_images Yes Yes
generate_stock_research_report Yes Yes

MCP Transport

All official MCP tools have REST SDK coverage. Use an MCP-compatible client (e.g. Claude, or any MCP host) when you specifically need MCP transport, tool discovery, or assistant-native tool calls — see the MCP server docs for setup.

Scope

This package is a REST API wrapper only. It does not implement MCP transport, SSE, streaming, tool discovery, Claude integration, or proprietary finance logic.

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

Links

License

MIT. See LICENSE.

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

hpsilab_mcp-0.7.0.tar.gz (23.7 kB view details)

Uploaded Source

Built Distribution

If you're not sure about the file name format, learn more about wheel file names.

hpsilab_mcp-0.7.0-py3-none-any.whl (15.1 kB view details)

Uploaded Python 3

File details

Details for the file hpsilab_mcp-0.7.0.tar.gz.

File metadata

  • Download URL: hpsilab_mcp-0.7.0.tar.gz
  • Upload date:
  • Size: 23.7 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/6.2.0 CPython/3.12.4

File hashes

Hashes for hpsilab_mcp-0.7.0.tar.gz
Algorithm Hash digest
SHA256 ac431adf49606d31c81f73e2676cbcddc21a20902d755c0a7392eefafe21744a
MD5 340081300d9d5d91933b6bba3fc4ee38
BLAKE2b-256 d81d367183ef5d8f970f4f588de290d2e0c19c14e45e857b86f9421fc6619111

See more details on using hashes here.

File details

Details for the file hpsilab_mcp-0.7.0-py3-none-any.whl.

File metadata

  • Download URL: hpsilab_mcp-0.7.0-py3-none-any.whl
  • Upload date:
  • Size: 15.1 kB
  • Tags: Python 3
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/6.2.0 CPython/3.12.4

File hashes

Hashes for hpsilab_mcp-0.7.0-py3-none-any.whl
Algorithm Hash digest
SHA256 df64e6e134304799888e40442f7da7558bd54c3957e0481511ce597346abe2de
MD5 aae3d9ce0bbea4041c3756bf263639d1
BLAKE2b-256 1fafadc2841be4718dacde371402c1774917650b8a17f8b05f479660f1a39da2

See more details on using hashes here.

Supported by

AWS Cloud computing and Security Sponsor Datadog Monitoring Depot Continuous Integration Fastly CDN Google Download Analytics Pingdom Monitoring Sentry Error logging StatusPage Status page